WO1985000453A1 - Data processing system - Google Patents

Abstract

In order to increase flexibility and decrease cost in a micro-code controlled environment through the use of pipeline processing, a data processing system is provided for processing a sequence of program instructions having two independent pipelines: an instruction pipeline (2, 3, 4) for reading instructions from storage and providing address data for the execution pipeline, and an execution pipeline (5, 6, 7) for referencing stored data via the addresses provided by the instruction execution, each having a plurality of serially operating stages (2, 3, 4, 5, 6, 7). Both pipelines operate synchronously under the control of a pipeline control unit (1) which initiates operation of at least one stage of the execution pipeline prior to completion of operations in the instruction pipeline, providing an operation overlap of one stage of each of the pipelines for each particular instruction. The pipeline control unit (1) can independently control the flow of instructions through the pipelines allowing conditional branching and subroutine operation.

Description

DATA PROCESSING SYSTEM

BACKGROUND OF THE INVENTION

The present invention relates to the field of digital computers and, in particular, to apparatus and methods for processing instructions in high speed data processing systems.

Data processing systems generally include a central processor, an associated storage system (or main memory) , and peripheral devices and associated interfaces. Typically, the main memory consists of relatively low cost, high-capacity digital storage devices. The peripheral devices may be, for example, non-volatile semi-permanent storage media, such as magnetic disks and magnetic tape drives. In order to carry out tasks, the central processor of such systems executes a succession of instructions which operate on data. The succession of instructions and the data those instructions reference are referred to as a program.

In operation of such systems, programs are initially brought to an intermediate storage area, usually in the main memory. The central processor may then interface directly to the main memory to execute the stored program. However, this procedure places limitations on performance due principally to the rela¬ tively long times required in accessing that main memory. To overcome these limitations a high speed (i.e. relatively fast access) storage system, in some cases called a cache, is used for holding currently

used portions of programs within the central processor itself. The cache interfaces with main memory through memory control hardware which handles program transfers between the central processor, main memory and the peripheral device interfaces.

> One form of computer, typically a "mainframe" computer has been developed in the prior art to con¬ currently hardware process a succession of instructions in a so-called "pipeline" processor. In such pipeline processors each instruction is executed in part at each of a succession of stages. After the instruction has been processed at each of the stages, the execution is complete. With this configuration, as an instruction is passed from one stage to the next, that instruction is replaced by the next instruction in the program. Thus, the stages together form a "pipeline" which, at any given time, is executing, in part, a succession of instructions. Such instruction pipelines for pro¬ cessing a plurality of instructions in parallel are found in several mainframe computers. These processors consist of single pipelines of varying length and employ hardwired logic for all data manipulation. The large quantity of control logic in such machines makes them extremely fast, but also very expensive.

Another form of computer system, typically a

"minicomputer," incorporates microcode control of instruction execution. Generally, under microcode control, each instruction is fully executed before exe¬ cution of the next instruction begins. Microcode- controlled execution does not provide as high perfor¬ mance (principally in terms of speed) as hardwired control, but the microcode control does permit signifi¬ cant cost advantages compared to hardwired systems. As a result, microcode control of instruction execution

OMPI _u has been employed in many cost-sensitive machines. Microcode reduces the total quantity of hardware in the processor and also allows much more flexibility in terms of adapting to changes which may be required during system operation. Unfortunately, the conven¬ tional pipeline techniques for instruction execution a e ^'not compatible with the multiple steps which must be performed to execute some instructions in a icrocode-controlled environment.

Accordingly, it is an object of the present invention to provide an improved computer system.

Another object is to provide performance characteristics heretofore associated only with mainframes while maintaining a cost profile consistent with the minicomputers.

It is yet^" another object to provide a com¬ puter system incorporating pipelined instruction pro¬ cessing and microcode-controlled instruction execution.

SUMMARY OF THE INVENTION

The invention relates to a data processing system and pipeline control method for processing a sequence of program instructions in a computer. The data processing system has an instruction pipeline having a plurality of serially operating instruction stages for reading instructions from storage and for forming therefrom plural address data to be employed during execution of the program instructions. The data processing system further has an execution pipeline having a plurality of serially operating execution stages for receiving the address data and for employing that data, formed by the instruction pipeline for

referencing stored data to be employed for executing the program instructions.

The data processing system features a pipe¬ line control unit for synchronously operating the instruction pipeline and the execution pipeline. The pipeline control unit has circuitry for initiating operation of at least one stage of the execution pipe¬ line using at least one of the address data formed by the instruction pipeline for program instruction prior to the completion of address data formation by the instruction pipeline for that program instruction. Thereby, operation of at least one instruction stage and one execution stage of the respective pipelines overlaps for each program instruction.

The data processing system further features sharing a memory between the instruction pipeline and the execution pipeline. A pipeline master clock for timing the pipeline stages has at least two clocked periods allotted for each stage of the pipeline to complete its operation. During one of these two clocked periods the instruction pipeline has access to the high speed memory and during another one of the clocked periods the execution pipeline has access to the high speed memory.

The pipeline control unit further has cir¬ cuitry responsive to exception conditions on the execu¬ tion and instruction pipelines for independently controlling, for each pipeline, the flow of instruction operations through the execution and instruction pipe- lines. Flow control of the instructions can include halting one or the other, or both, of the execution and instruction pipelines; running .the execution pipe-, line using artificial "NOP" (no operation) instructions

O PI while a previously empty instruction pipeline is being filled; extending the time for all pipeline stages to complete an operation for allowing one of the stages to complete its operations; providing extended time for a plurality of microinstructions to be used in the execu¬ tion stages of the pipeline; maintaining the instruc¬ tion pipeline in a halted state; and similarly related delay type operations.

The data processing system further features circuitry for detecting collisions between pipeline stages and for halting operation of one or more of the pipeline stages to provide a separation between the colliding instructions. Similarly, where a latter stage of the pipeline alters the data to be used by an earlier stage, an appropriate "flushing" of the pipe¬ line can be accomplished should a previously employed address be incorrect, or the instruction pipeline can be halted until the data is properly available from the execution pipeline (registered by-pass).

In another aspect of the invention, a pipe¬ line control method for use with an instruction and an execution pipeline having a plurality of serially operating instruction stages, features the steps of synchronously operating the instruction and execution pipelines and initiating operation of at least one stage of the execution pipeline using address data formed by the instruction pipeline at a time prior to completing, by the instruction pipeline, generation of all of the address data formed for a particular program instruction.

In another aspect, the data processing system relates to an instruction pipeline and an execution pipeline each having a plurality of serially operating

OMPI ^• instruction .stages and a pipeline control unit for operating the instruction and execution pipelines. In this aspect of the invention there is featured cir¬ cuitry for detecting collisions between read data from a register in the instruction pipeline phase of opera¬ tion, read in response to a first program instruction, and write data written in registers during the execu¬ tion phase of operation in response to an earlier instruction. During the execution phase, a plurality of execution cycles, during each of which a register can be modified, occur, and the first instruction will require one of the modified values to continue valid operation. The detection circuitry features storage for the modified values generated during the execution phase and storage for the write register addresses associated therewith. Circuitry is provided for com¬ paring the associated write register address for each modified value with the read register address employed by the instruction pipeline. When a match is found, directing circuitry replaces the modified value, to be written at the matched address for the data previously designated to be used during the associated instruction phase of operation.

In selected embodiments of this aspect of the invention, only a portion of the originally read data may be modified and replaced. In other aspects of the invention, operation of all pipeline stages, except those required to generate modified data, can take place so that the modified data can be generated prior to a continuation of operation along the instruction, and portions of the execution, pipelines.

The invention also features circuitry for detecting potential collisions between data read from a register in the instruction pipeline phase of operation

B-7'*5" -*?""_ 'T^* 7?T and data which may be written at a later time. According to this aspect of the invention, the instruc¬ tions themselves are examined and potential collisions predicted. In response to the prediction of a poten- tial conflict, the pipeline control unit provides delay between the potentially colliding instructions so that if a collision takes place, it can be handled in due course by the system structure described above.

In yet another aspect of the invention, the invention relates to a program instruction flow predic¬ tion apparatus and method for a data processing system having means for prefetching an instruction. The flow prediction apparatus features a flow prediction storage element, an instruction storage element containing program instructions to be executed, circuitry for addressing an instruction in the instruction storage element, containing program instructions circuitry for addressing a flow control word from the flow prediction storage element at a location derived from the location of the addressed instruction in the instruction storage element. The flow control word contains at least a branch control portion for predicting the flow of program instructions and a next program instruction address portion containing at least a portion of a next program address if a program branch is predicted.

In a particular embodiment of the invention, the flow prediction storage element is a random access, high speed memory. The flow prediction memory has a significantly smaller storage capacity than the system main memory and the flow prediction storage element can have fewer storage locations than the instruction storage element.

In particular, the flow prediction apparatus further features monitoring circuitry for monitoring

- . _02-/PI __ instruction flow during the instruction prefetch and circuitry responsive to the monitoring circuitry for updating the flow prediction storage element based solely upon a history of the instruction flow. Specifically, the monitoring circuitry can respond to only the instruction flow of a most recent execution of the present instruction. In other aspects, the program instruction flow prediction apparatus features cir¬ cuitry for employing a flow altering data from the flow prediction storage element in place of a next sequen¬ tial flow data during a next program instruction fetch operation.

According to the program instruction flow prediction method, there are featured the steps of addressing an instruction in an instruction storage element, reading a flow control word from a flow pre¬ diction storage element at a location derived from the location of the addressed instruction in the instruc¬ tion storage element, the flow control word containing at least the branch control portion and the next program instruction address portion, and altering the otherwise predetermined instruction flow in accordance with said program instruction address portion if a program branch is predicted.

BRIEF DESCRIPTION OF THE DRAWINGS

The foregoing and other objects of this invention, the various features thereof, as well as the invention itself, may be more fully understood from the following description, when read together with the accompanying drawings in which:

Fig. 1 shows, in block diagram form, an exemplary computer system embodying the present inven¬ tion.

OMPI Fig. 1A depicts, in block diagram form, the instruction processor, including the two three-stage pipelines, showing overlap and flow between stages, and the pipeline control unit, of the central processor of the system of Fig. 1;

^* Fig. 2 depicts the five hardware units that form the instruction processor of Fig. 2, showing major data paths for the processing of instructions;

Fig. 3 shows, in block diagram form, the pipeline control unit of Figure 2;

Fig. 3A shows, in block diagram form, the decode logic for the pipeline control unit of Fig. 4;

Fig. 4 shows, in detailed block diagram form, the pipelines of Fig; 1;

Fig. 5 depicts the flow of instructions through the two pipelines, with examples of alteration to normal processing flow;

Fig._.6 illustrates the clock generation of the ID stage of the IP pipelines of Fig. 1A;

Fig. 7 depicts a block diagram of the Shared

Program Cache of Fig. 1A;

Fig. 8 depicts a block diagram of the Instruction Pre-Processor of Fig. 1A;

Fig. 9 depicts a block diagram of the Micro-Control Store of Fig. 1A; and

Fig. 10 depicts a combined block diagram of the two Execution units of Fig. l_.A.

OM?I Fig . 11 shows , in block diagram form, the branch cache of the system of Fig . 4 ; and

Fig . 12 shows , in block diagram form, the register bypass network of the Instruction Pre-Processor of Fig . 8.

>

DESCRIPTION OF THE PREFERRED EMBODIMENT

Fig. 1 shows a computer system embodying the present invention. The system includes a central pro¬ cessor, main memory, peripheral interface and exemplary peripheral devices.

This system of Fig. 1 processes computer data instructions in the central processor which includes instruction pre-processing hardware, local program storage, micro-control store, and execution hardware. The central processor includes two independent pipe¬ lines; the Instruction Pipeline (IP) and the Execution Pipeline (EP). In the preferred form, each pipeline is three stages in length (where the processing time asso¬ ciated with each stage is nominally the same), with the last stage of the IP being overlapped with the first stage of the EP. With this configuration, an instruc¬ tion requires a minimum of five stage times for comple¬ tion. All control for advancing instructions through all required stages originates from a Pipeline Control Unit (PCU) in the central processor. The PCU controls the stages to be clocked dynamically, based on pipeline status information gathered from all stages.

This form of the invention processes instruc¬ tions defined in the System Architecture Reference Guide, 2d Ed. (PRC3060-182) Revision 18,2, published by Prime Computer, Inc., Natick, Massachusetts, and sup-

ports the machine architecture, which includes a plura¬ lity of addressing modes, defined in the Reference Guide. In keeping with this architecture, words are 16 bits in length, and double words are 32 bits in length. This form of the invention is optimized to perform address formations including BR + X + D, BR + GRH + D and RP + X + D, where BR (Base Register) is a 32-bit starting address pointer, X (Index) is a 16-bit register, GRH (high side of General Register) is a 16- bit quantity, D (the displacement) is contained expli¬ citly in the instruction and may be either 9 or 16 bits, and RP is the current value of the program counter.

PRINCIPLES OF PIPELINE OPERATION

Pipeline Stage

Fig. 1A shows, in functional block diagram form, a three-stage pipelines an Instruction Pipeline (IP) and an Execution Pipeline (EP) , together with the pipeline control unit (PCU) in the central processor. The Instruction Pipeline includes an Instruction Fetch (IF) stage 2, an Instruction Decode (ID) stage 3, and an Address Generation (AG) Stage 4. The Execution Pipeline (EP) includes a Control Formation (CF) stage 5, an Operand Execute (OE) stage 6, and an Execute Store (ES) stage 7. The PCU 1 is depicted in detailed block diagram form in Figs. 3 and 3A and the IF, ID, AG, CF, OE and AS stages are depicted in detailed block diagram form in Fig. 4.

Fig. 2 shows an embodiment of the IP, EP and PCU of Fig. 1A in terms of five hardware units:

Instruction Pre-Processor (IPP) 8, Shared Program Cache (SPC) 9, Execution board-1 (EXl) 10, Execution-2 board

(EX2) 11, and Mirco-Control Store (MCS) 12. The hard¬ ware units of Fig. 1A are representative of groupings of the various elements of the IP and EP of Fig. 4. The respective hardware units are shown in detailed form in Figs. 7-10. In alternative embodiments, other groupings of the various elements of the IP and EP may be used.

Briefly, in the illustrated grouping of Fig. 2, the Shared Program Cache 9 contains local storage and provides instructions by way of bus 13 to the Instruction Pre-Processor 8, and provides memory operands by way of bus 14 to the Execution 1 board 10. The IPP 8 supplies memory operand addresses by way of bus 15 to the SPC 9, register operands and immediate data by way of bus 17 to EXl 10, and control decode addresses by way of bus 19 to the Micro-Control Store 12. EXl 10 operates on memory operands received by way of bus 14 from the SPC 9 and register file operands received by way of bus 16-> from the Execution 2 board 11, and transfers partial results by way of bus 18 to EX2 11 for post-processing and storage. EX2 11 also performs multiplication operations. The MCS 12 provi¬ des microprogrammed algorithmic control for the four blocks 8-11, while the PCU 1 provides pipeline stage manipulation for all blocks 8-12.

The pipeline stage operations are completed within the various hardware units 8-12 as follows:

IF (Instruction Fetch): A Look-ahead program counter on SPC 9 is loaded into a local general address register; instruction(s) are accessed from a high speed local memory (cache) . ID (Instruction Decode): Instruction data is transferred from SPC to IPP 8; IPP 8 decodes instructions, forming micro-control store entry point information for MCS 12, and accessing registers for address generation in

IPP 8.

AG (Address Generation): IPP 8 forms instruction operand address and transfers value to SPC 9 address register.

CF (Control Formation): MCS 12 accesses local control store word and distributes control information to all boards.

OE (Operand Execute) : SPC 9 accesses memory data operands in cache; EXl 10 receives memory data operands from SPC 9, register operands from IPP 8, and begins arithmetic operations.

ES (Execute Store): EXl 10 and EX2 11 complete arthimetic operation and store results.

The Address Generation and Control Formation stages are overlapped in time within the data system. The IP and EP operate synchronously under the supervision of the pipeline control unit (PCU) 1, which interfaces to each stage with two enable lines (ENCxxl and ENCxx2) that provide two distinct clock phases within each stage, as indicated in Fig. 1A. The notation "xx" refers to a respective one of the reference designations IF, ID, AG, CF, OE and ES. The six ENCxx2 lines denote the respective stage operations are complete and the data (or control) processed in those stages are ready for passing to the next stage.

Clocking of .Pipeline Stages

Timing and clocking in the dual pipelines (IP and EP) are synchronized by two signals - the master clock MCLK and the enable-end-of-phase signal ENEOP. ENEOP is produced by the Pipeline Control Unit 1 and notifies all boards of the proper time to examine the stage clock enable signal lines (ENCxxl and ENCxx2) in order to produce phase 1 and phase 2 stage clocks from the master clock MCLK. (See Fig. 6). Pipeline stages always consist of two phases. Phase 1 lasts for exactly two MCLK pulses while phase 2 can last for an arbitrary number of MCLK pulses, as described below, depending on the conditions present in both the IP and the EP.

An example of how MCLK and ENEOP and the stage clock enables interact on each board to form the clocks which define the stage boundaries is shown in Fig. 6 for the Instruction Decode stage 2. Register 22 generates clock signals when enabled by ENEOP. When ENCIDl is present the clock CIDl is generated; when ENCID2 is present, the clock CID2 is generated.

PIPELINE CONTROL UNIT

The Pipeline Control Unit 1 shown in Figs. 3 and 3A controls the flow of instructions through the dual pipelines (IP and EP) by generating the enable signals for all clocks which define stage boundaries and relative overlap of the IP and EP. - The PCU 1 includes stage clock enable decode logic 23 and the Pipeline State Register (PSR) 24. PCU 1 receives as inputs:

1. Instruction information and exception and register conditions from the IPP 8

SUE 2.. Exception and cache conditions from the SPC 9

3. Microcode specified timing conditions related to the length of stage OE and the overlap of stage OE and CF from the MCS 12

»

4. Exception conditions from EXl 10 and EX2 11.

The PCU 1 has complete control of all stage boundaries. With that control:

1. The PCU 1 can hold the IP while cycling multi-microcode through the EP.

2. The PCU 1 can alter the flow of instruc¬ tions based on control information provided by microcode.

3. The PCU 1 can extend all stages if extra time is required for a particular stage to finish its operation.

4. The PCU 1 can alter the relative overlap of stages OE and CF of the EP in order to allow different types of microcode sequencing

(as described below in conjunction with EX1,2).

5. The PCU 1 can flush out instructions in the IP and recycle the IP to load new instructions upon detecting incorrect flow

(such as an incorrect flow prediction pro¬ vided by Branch Cache 34).

SUBSTITUTE

6.. The PCU 1 can idle the EP with no- operation (NOP) cycles, while cycling the IP, for example, when.IRP 27,33 in the SPC 9 is reloaded after an incorrect program flow sequence.

^* 7. The PCU 1 can suspend all pipeline opera¬ tions during non-overlappable operations such as "cache miss" access to main memory.

8. The PCU 1 can introduce separation bet- ween sequential instructions in the IP under certain conditions, such as "collisions" bet¬ ween instructions.

9. The PCU 1 can keep an instruction held in the IF stage upon detecting an instruction- related exception, and then allow the other instructions currently in the pipeline to complete processing so that the exception can be processed in the correct order.

The Pipeline Control Unit (PCU) 1 which controls the clocking of the stages in the IP and EP is shown in detail in Fig. 3A. Condition signals received from the IPP 8, SPC 9, MSC 12, EXl 10, and EX2 11 hard¬ ware units are utilized to produce enable signals for clocks in the IF 2, ID 3, AG 4, CF 5, OE 6, and ES 7 stages of the dual pipelines (IP and EP) . There are two major elements in PCU 1 which produce the clock enable signals ENCxxl,2: a pipeline state register (PSR) 24 (including state registers 180,182,184,186,188,190) and a stage clock enable decode logic 23 (including conbinatorial logic blocks (181,183,185,187,189,191). The state registers

180,182,184,186,188,190 indicate that the respective

C PI pipeline stages are ready to be enabled. if there are no conditions received by the PCU 1 which should inhi¬ bit the stage from proceeding. When the stages are in operation, the state registers 180,182,184,186,188,190 provide a timing reference to distinguish between the two phases of each stage. The combinatorial logic blocks 181,183,185,187,189,191 decode the conditions received from the various hardware units 8-11 to deter¬ mine whether or not the stage operation should proceed.

The values of the state registers are controlled by the various ENCxxl and ENCxx2 signals as follows:

The IF state register IFSR 180 is set ready by ENCIF2 which indicates that an instruction fetch is complete and another can begin.

ENCIF1 sets state register IFSR 180 to indi¬ cate that phase 1 of the IF stage has been performed.

The ID state register IDSR 182 is set ready by ENCIF2 which indicates that the

IP prefetched an instruction which is ready to be decoded. ENCIDl sets state register IDSR 180 to indicate that phase 1 of the ID stage has been performed.

The AG state register AGSR 184 is set ready by ENCID2 which indicates that the IP has decoded an instruction which now requires an operand address generation. ENCAG1 sets state register AGSR 184 to indicate that phase 1 of the AG stage has been performed.

The CF state register CFSR 186 is set ready by ENCCF2 which indicates that the EP has

#»ι • ITS ,"•"!^■" 3 "W'? ST""? rr* -7T I _> O ^* PI ^{* * *} completed formation of the control word asso¬ ciated with the microinstruction ready to enter the OE stage. ENCCFl sets state register CFSR 186 to indicate that phase 1 of the CF stage is complete.

^* The OE state register OESR la^'s is set ready by ENCCF2 which indicates that control and addressing information is ready to be passed to the OE stage. ENCOE1 sets state register OESR 188 to indicate that phase 1 of the OE stage is complete.

The ES state register ESSR 190 is set ready by ENCOE2 which indicates that operands are ready to enter the final execution stage and be stored. ENCES1 sets state register ESSR

190 to indicate that phase 1 of the ES stage is complete.

Combinatorial logic networks EN1F 181, ENID 183, ENAG 185, ENCF 187, ENOE 189, and ENES 191 monitor condition signals received from the hardware units 8- 11, and when those conditions indicate, block the ENCxxl and ENCxx2 enables for the respective stages. In Fig. 3.A, each signal entering the combinatorial logic blocks may inhibit the respective enables for that stage. The condition signals applied to the PCU 1 are described below.

The IPP 8 provides two conditions signals to the PCU 1: COLPRED and COLDET. COLPRED (collision predicted) indicates that separation may have to be introduced between two instructions in the IP to allow determination of whether or not a register collision exists. COLPRED holds the IF, ID, and AG stages of the

"!!^*'''^,,°' s"^*"-"^**- tr-*'"^{* *} ' i IP to permit determination of whether or not a register collision exists between the instruction in the ID stage and the instruction that has just entered the EP. Logic ENID 183 generates FORCENOP (force a no operation instruction in the CF stage) , when no new instruction is available to enter the EP. This signal disables the LOA signal on bus 91 by setting LDA register 84 to zero. COLDET indicates that a collision does exist. IN response, the generation of the clock enable signal for stages IF, ID, AG, CF, and OE is delayed until the updated register is available from the completion of the ES stage. This process is illustrated in Fig. 5 during time periods T24, T25, and T26.

SPC 9 provides three condition signals to PCU 1: CACHEMISS, IMEMEXCPTN, OPMEMEXCPTN. CACHEMISS indicates that a cache miss has occured in the SPC 9. In response to the CACHEMISS signal, the generation of the clock enable signals for the stages IF, 10, AG, CF, and OE is delayed until the memory subsystem has updated the cache. The signal IMEMEXCPTN from the SPC 9 indicates that an exception (such as an access viola¬ tion, STLB miss) has occurred during an instruction fetch. The IMEMEXCPTN signal similarly effectively holds the IF stage from further prefetching and pre- vents the instruction in the IF stage from proceeding to the ID stage. All other stages are allowed to pro¬ cess, so that the pipeline may be emptied of all instructions before proceeding to handle the exception condition. The OPMEMEXCPTN signal indicates that an exception has occurred during the operand fetch in stage OE. This OPEMEMEXCPTN signal blocks stages IF, ID, AG of the IP and provides sufficient delay for the CF stage as to allow the EP to branch to a microcode routine capable of handling the exception condition. Stage OE, in which the exception occurred, is effec¬ tively cancelled.

OMPI The MCS 12 provides information decoded from microcode related to the number of microcode-driven execution cycles required to complete an instruction and the timing required for completing data manipula- tion and formation of micro-control store addresses within such cycles. Three signals within this category are produced. EXCMPL is only asserted on final microsteps of instructions. During all other microsteps of instructions, the PUC 1 holds the IP con- sisting of stages IF, ID and AG until the multi- microcode has completed. XTNDEX indicates that additional time is required in the OE stage, while XTNDCTRL controls the relative overlap of stages OE and CF, allowing microcode jump conditions -to be used in the present microstep to select following microstep.

The MCS 12 also produces FLUSH in cases where incorrect instruction flow has occurred, such as when wrong branch cache predictions are made. In response to the FLUSH signal, all IP stages are cleared and a new IF stage is started.

The EXl,2 pair 10,11 produces the signals EXECEXCPN, which is generated under certain execution- related conditions, and CEXCMPL, which indicates whether or not a microinstruction is a final one based on testing bits within EX1,2 10,11. In response to EXECEXCPN, the PCU 1 functions in a similar manner as in response to OPMEMEXCPTN, differing only in the microcode routine which is executed. The CEXCMPL causes the same result as EXCMPL, differing -only in that the generation of CEXCMPL is conditioned on cer¬ tain test bits within EX1,2 10,11.

OMPI ' INSTRUCTION FLOW IN PIPELINES

Fig. 5 shows the flow on instructions through the six stages of the dual pipeline (IP and EP) , and shows the clocking associated with those stages. In Fig. 5, Tl .- T27 are time reference markers; II - 125 represent machine instructions; Ml - M6 represent addi¬ tional microcode execution cycles required to complete the execution of a machine instruction and N represents a NOP (or "no-operation") instruction cycling through the Execution Pipeline.

Time periods Tl and T2 show the dual pipe¬ lines concurrently processing five machine instruc¬ tions. Instruction 4 requires an additional microcode cycle (Ml); during time period T3, the PCU 1 idles the IF, ID, and AG stages of the Instruction Pipeline.

During T4, the IP again begins to advance instructions. 15 also requires an extra execution cycle (M2), so that during time periods T5 and T6, the PCU 1 again idles the three stages of the IP. The second microcode step for 15 (i.e. M2) is conditional, based on the results of the execution of 15; the PCU 1 therefore stretches the CF stage for M2 relative to the end of the OE stage for 15. Both pipelines are operative again during time periods T7 and T8. 17 is an example of a machine instruction requiring four extra microcode execution cycles (M3, M4, M5, and M6). The PCU 1 begins and con¬ tinues to idle stages IF, ID, and AG beginning in time period T9. Microcode execution cycle M3 requires addi¬ tional time in the OE stage, so the PCU 1 extends both the CF and OE stage from T10 to Til.

In the exemplary sequence of Fig. 5, 17 is a conditional instruction. During the multiple cycles of execution associated with 17 (i.e. M3 - M6), the system determines that the IP has prefetched incorrectly. The EP then flushes the pipeline by notifying the PCU and reloading the look-ahead program counter used for prefetching. The IF, ID, and AG sta- ges of the Instruction Pipeline are shown refilling during time periods T14, T15, and T16. While the IP is refilling, the EP completes the last microcode step associated with 17. During time periods T14 and T15, NOP steps are forced into the Execution Pipeline, as no machine instruction is yet available for execution.

118 is an example of a machine instruction requiring extra time in the OE stage. The PCU also ^* delays the IF, ID, AG, and CF stages of the instruc¬ tions behind 118 (i.e. 119, 120, and 121) keeping all s^'tages in synchrony.

Time periods T23, T24, T25, and T26 show an example where the IP requests special action in the PCU prior to advancing 122 from the ID stage to the AG stage. In particular, the IP has determined that 121 will modify a register required by 122 to generate the operand address associated with 122► In response, the PCU 1 suspends the IP during time period T24, and delays the IF, ID, and AG stages in the IP and the CF stage in the EP during time periods T25 and T26, so that the results stored for 121 in the ES stage can be used by the AG stage for 122. Because no machine instruction is available at time period T24, a NOP cycle is introduced into the CF stage of the EP.

The phased stage clocks (Cxxl,Cxx2) described in the Pipeline Control Unit section are shown beneath the instruction flow diagram in Fig. 5.

PIPELINE ELEMENTS

As described above. Fig. 4 shows the prin¬ ciple hardware elements contained in each of the six stages of the instruction and execution pipelines. In the embodiment of Fig. 2, several of the stages include elements which are time-multiplexed resources within the pipelines. These elements are shown with identical references designations in the various stages of the Fig. 4 configuration.

For a single machine instruction passing through the pipeline stages, the processing occurring within the IF stage is confined to hardware on the SPC 9. During the first phase of the IF stage, the con¬ tents of the look-ahead program' counter 27,33 are gated through the SPC's address selector 29,39 and loaded into the address registers 44,40 with clock pulse CIF1. During the second phase, 32 bits of instruction data are retrieved from cache 41 and loaded into the cache data register 42 with clock pulse CIF2, which ter- minates the IF stage. The STLB 45 is also accessed during the second phase, loading a mapped physical memory address into register BPMA 46 for possible use in the event data is not contained in cache 45. The branch cache 34 is also checked during the IF stage. As described below in conjunction with Fig. 11, based on the information contained, register IRP 27,33 is either loaded with a new target address or incremented.

During the first phase of the ID stage, the instruction data held in the cache data register 42 is passed through selectors 47,33 on the SPC 9 ensuring that the opcode for the instruction at the current program counter value is presented on bus 63. The thirty two bits of instruction data are passed on buses

OMPI 62,63 to the opcode latches and selectors 80,81 on the IPP 8; this data is retained on the IPP 8 by clock pulse CIDl. During the later phase of the ID stage, opcode information is used to access the microcode entry point for the instruction from the decode net 82 which is loaded into register LDA 84 with clock pulse CID2. A'lso during the second phase, registers required for memory address generation are accessed from register file AGRF 72 and stored in register BXR 73 with clock pulse CID2. Finally, the displacement required for address generation is transferred from the instruction latches and selectors 80,81,207 and loaded into the^" pipeline displacement register DISP 83 through selector 209 with clock pulse CID2. Summarizing, at the end of the ID stage, information for the CF stage and AG stage has been stored in pipeline registers; the machine instruction processing then simultaneously moves into the last (AG) stage of the Instruction Pipeline and the first (CF) stage of the Execution. Pipeline.

During the AG stage, the IPP 8 computes the effective address of the memory operand (assuming the instruction being processed requires a memory reference) and loads that address into the address registers on the SPC 9. The operation commences with a selector 74 choosing either the output of register BXR 73, which contains the contents fo the appropriate ■ registers accessed during the ID stage, or BDR71 which contains an updated value of a register (as described in detail below with respect to register bypassing in the IPP section) . The first ALU 75 then adds the base register and index register as specified by the instruction and feeds the value into the second ALU 76 where it is combined with displacement offset from register DISP 83. The resulting operand address is passed through selectors 86,78 and sent to the SPC 9 on

^' buses 49,57. Selectors 28,39 on the SPC 9 gate the address to the cache 41 and STLB 45 through address registers 44,40 which are loaded with clock pulse CAG2. A copy of this address is also stored in the IPP 8 in registers EAS 85,77 for later use if the particular machine instruction requires multiple microcode execu¬ tion cycles.

The CF stage performs the access and distri¬ bution of the micro-control store word used for algorithmic control to all hardware units. In the case of a machine level instruction, the entry point from the ID stage is chosen by the selector 103 and pre¬ sented to the micro-store 104. The output of the microstore is driven to all required hardware units through buffer 105 and loaded into a plurality of control word registers 215,65,216,145 with clock pulse CCF2, which marks the end of the CF stage. Also at the end of the stage, the current microstore address is loaded into the holding register RCH 106 with clock pulse CCF2.

The ALU 118 operation completes during the first phase of the ES stage; ALU data is passed through selectors 119,121 for post processing, including shifting, and loaded into registers RD 122 and RS 126 with clock pulse CES1. Finally during the last phase of the pipeline, results of the calculation stored in register RS 126 are written into register file 130 if so specified by the micro-store control word and into register BDR 71 clocked at CES2. Register BDR 71 makes an updated location available to hardware in the ID stage for updating register file AGRF 72 and for bypassing AGRF 72 in calculating an operand address in the AG stage through selector 74.

SUBSTITUTESHEET

' O PI In certain cases, a particular machine instruction will require more than one cycle in the EP. In such a case, the PCU 1 will stop providing clock enables to the IP, but continue to cycle the three sta- ges in the EP. The micro-store 104 permits any general purpose algorithm to execute within the EP. Results computed in the OE and ES stages and loaded into registers RD 122 and RS 126 with clock pulse CES1 can be fed back into the ALU 118 via the ALU selectors 117,125, thus enabling data manipulation in successive execution cycles to also be pipelines. In the event that an execution cycle references a register written in the previous cycle, the value in register RS 126, which will be written into the register file 130 during the last phase of the ES stage, can bypass register RI 129 normally used to read register file data and be presented directly to selector 125 and presented to the ALU 118.

DESCRIPTION OF THE HARDWARE UNITS 8-11

Shared Program Cache

The Shared Program Cache 9 in Fig. 7 includes a high speed cache memory 41 for instructions and operands, a segment table look-aside buffer (STLB) 45 for retrieving recently used mapped physical memory addresses, and a branch cache 34 used to predict the flow of conditional machine instructions as they are fetched from cache. Also shown are pipeline address and data registers used in conjunction with the storage elements.

In operation, the SPC 9 operates under the general control of enables from PCU 1, and, during the OE stage, also under the general control of microcode

stored in MCS 12, which has been transferred by way of RCC bus 64 to RCM register 65. Selectors 28,39 deter¬ mine the source for main SPC address busses 53,59 which load address registers 44,40 which in turn directly address the cache 41 and STLB 45. Also loaded from the main address buses 53,59 are backup address registers E.RMAH, ERMAL 30,37 for operand addresses and PRMAL 36 for the low side of the program counter. Backup address registers 30,37 provide backup storage of the cache and STLB addresses for use when the contents of the registers 40,44 (which directly access each 41 and STLB 45) are overwritten with new addresses prior to detection of a cache miss or memory exception.

There are four sources of addresses for accessing the cache and STLB storage elements: (i) registers IRPH 27 and IRPL 33 which contain the look- ahead program counter used for prefetching instruc¬ tions, (ii) buses BEMAH 49 and BEMAL 57 which transfer effective addresses generated in the IPP 8, (iii) buses BDH 50 and BDL 54 through buffers 26,31 which transfer addresses from EX2 11 during multiple micro¬ code sequences, and (iv) buses 51 and 56 which are used to restore addresses from the program counter backup registers 27,36 or operand address backup registers 30,37 previously used in the event of cache misses or memory exception conditions. Thirty-two bits of information from cache 41 are stored in a data register 42 and gated on bus 60 to selectors 43,47, from which data is driven to EXl 10 and instructions are sent to the IPP 8 over buses BBH and BBI. 63,62.

In the event of cache misses or explicit main memory requests, virtually mapped physical addresses from the STLB 45 or absolute addresses from the backup registers 27,30 and 36,37 are gated to selector 46 and stored in the BPMA register 48. The physical memory address is then fed through selector 47 and gated on to BBH, BBL 63,62 and transferred to the main memory sub¬ system. The backup registers 27,36 and 30,37 are also selectively transferred to EXl 10 over buses BBH, BBL 63,62 for fault processing through the appropriate selectors 29,38,47,43.

The branch cache 34 permits non-sequential instruction prefetching based on past occurrences of branching. Briefly, the branch cache 41 is addressed by the low-side of the look-ahead program counter IRPL 33; the output from that operation consists of control information indicating whether o not to re-load IRPL 33 with a new target address on bus 55 through selector 32. As described in detail below, the information in the branch cache 34 is maintained by the execution hardware and is updated along with IRPL 33 by way of bus BDL 54 whenever "it is determined (in IPP 8) that incorrect prefetching has occurred. In the event the branch cache 34 does not indicate that the prefetch flow should be altered, program counter- IRPL 33 is then incremented. When the branch cache 34 does alter program flow, the new contents of IRPL 34 are gated onto bus BEMAL 57 by way of buffer 35 and sent to the IPP 8 for variable branch target validation.

Instruction Pre-Processor -

The Instruction Pre-Processor (IPP) 8 shown in Fig. 8 includes instruction alignment logic, decoding hardware, arithmetic units for address genera- tion, and registers for preserving addresses trans¬ ferred to the SPC 9. The input logic of the IPP 8 is adapted to process one- and two-word instruction for¬ mats and to accommodate the instruction fetching in the

OMPI SPC -9 which is always aligned on an even two-word boun¬ dary. In either instruction format, the first word always contains the opcode and addressing information; for one-word instructions the displacement for address offset is also contained in the same word; for two-word instructions, the displacement is contained in the second word.

In instruction prefetching operation, the IPP 8 operates under the control of the enables received from PCU 1; during processing of multiple execution cycles, registers are updated and manipulated under the general control of microcode stored in MCS 12, which has been transferred by way of RCC bus 64 to RCM register 215. The SPC 9 transfers two words of instruction information to the IPP 8 over buses BBH 63 and BBL 62. The two words of instruction data pre¬ sented to the IPP 8 can be various combinations, such as two one-word instructions, an aligned (even boundary) two-word instruction, or the second word of a two-word instruction and the next one-word instruction. The SPC 9 gates the opcode of the instruction asso¬ ciated with the current value of the program counter IRPL 33 onto BBH 63 where it passes through the OPCL 80 selector latch for immediate processing.

The contents of BBL 62 are stored in register

IREG 81; depending on whether or not this second word contains an opcode or a displacement, the contents of IREG 81 is gated by way of bus 94 to the OPCL 80 latch, or to the selector 209. The output of the OPCL 80 latch is transferred by way of bus 93 to the decode net 82, the opcode register OPCR 207, the address inputs of register file AGRF 72 and register bypass blocks (including collision prediction logic 208 and collision detection logic 211). The decode net 82 provides control information for continuing the pre-processing of the instruction and also provides a micro-control store entry point which is stored in the LDA register 84 and subsequently driven to the MCS 12 over the bus LDA 91. The register bypass blocks are described in detail below.

Information decoded from the instruction governs if and how the operand address should be formed. Depending on whether an instruction contains one or two words, the selector 209 chooses either OPCR 207 on bus 203 or the IREG 81 on bus 94. If the instruction in stage IF is two words and unaligned, its displacement does not arrive from the SPC 9 until it has proceeded to stage ID.^" In this case, the DISP selector latch 83 selects a displacement value directly from bus BBL 62. Otherwise, latch 83 selects a displa¬ cement value from selector 209. The displacement value from latch 83 is coupled by way of bus 92 to the B-leg of ALU 76.

The IPP 8 includes a register file AGRF 72 which contains copies of all registers used in address calculation. The AGRF 72 can simultaneously access 32 bit base or general registers and 16 bit index registers transferring them into base and index pipe- line register 73. The true contents of these registers are maintained by the EX2 11 board in the execution unit and any changes to the registers do not occur until the ES stage of the execution pipeline. At the completion of stage ES, updated register contents are sent over BDH 50 and BDL 54 and through buffer 210 and are loaded into the bus D register BDR 71. The output bus 87 from BDR 71 distributes the contents of that register to the AGRF 72 (for updating register copies) and to the selector 74 (for register bypassing, as described in detail below, in conjunction with Fig. 12).

The collision detection logic 211 compares the AGRF 72 address (as decoded from the instruction in stage ID) to the address used by EX2 11 (as received in to the IPP 8 over bus BII 204) to write its register file. If the collision detection logic 211 determines that EX2 11 has updated a base, index or general register which matches the one just loaded from AGRF 72 into BXR 73, logic 211 selects the new register value held in BDR 71 in place of the output of BXR 73 by controlling selector 74.

Collision prediction logic 208 predicts possible collisions between instructions which are one stage apart in the IP by comparing the address being read from the AGRF 72 with a "guess" of a written address derived from bus 203. If. a possible collision is discovered, the PCU 1 is notified to separate the two instructions by one additional stage time so that the collision detection logic 211 can determine whether a problem actually exists. This technique of register bypassing is described more fully below.

As described fully below, selector 74 selec¬ tively gates the high word of the base or general register (as fetched from the AGRF 72) over bus 89 to selectors 212 and 86. The low word of the base or general register on bus 95 and the index register value on bus 95 are added together in the indexing ALU 75 if this operation is specified by the instruction. The displacement ALU 76 adds the result from the indexing ALU 75 to the displacement transferred from DISP 83 on bus 92. The result from ALU 76 is transferred to bus 90 to selectors 78 and 213 and to the branch cache validation logic 214. The branch cache validation logic 214 com¬ pares the computed branch address on bus 90 to the pre¬ dicted address from the branch cache 34 sent from the SPC 9 over bus BEMAL 57.

The effective address source registers (EASH

85 and EASL 77) and effective address destination registers (EADH 205 and EADL 206) function as two 32-bit memory address pointers, the low word of which (i.e. EASL 77 and EADL 206) are counters. EADH 205 and EADL 206 are loaded from bus 200. EASH 85 and EASL 77 are loaded from selector 212 over bus 201 and selector 213 over bus 202 respectively. Busses BBH 63 and BBL 62 are coupled to the outputs of selector 86 and 78 respectively, and provide general register and imme- iate operands to EXl 10. Busses BEMAH 49 and BEMAL 57 are similarly coupled to the output of selectors 86 and 78, respectively and provide memory addresses to the SPC 9 for referencing cache 41 and STLB 45. Data on busses 89 and 90 are transferred over busses BEMA 49,57 during stage AG of the IP by selectors 86 and 78.

During microcode controlled memory accesses, either EAS 85,77 or EAD 205,206 can be selected. Either EAS 85,77 or EAD 205,206 can also be selected onto busses 63,62 by selectors 86 and 78.

Micro-Control Store

The micro-control store unit 12 of Fig. 9 includes microcode storage 104, the next microcode address selector 103, the RBPA register 102, the pre¬ sent micro-address register RCH 106, the microcode stack 107, and the buffers 105 for driving new control bits (RCC's) by way of bus 64 to all boards.

The microstore 104 mau be selectively loaded to contain 5K 80 bit microcode words as provided over

OMPI ^' bus 108 from the BDH bus-50 by way of buffer 101. Of the 80 bits in each microcode word, 8 bits are directed to parity checking network 66, and the remaining 72 bits are transferred to the IPP 8, SPC 9, EXl 10 and

5 EX2 11 for algorithmic control during execution cycles. The microstore 104 and RCH 102 are addressed by way of bus 109. Bus 109 is driven by selector 103 which selects among the various sources for generating next addresses. These sources include the RBPA register 102

10 (which is used during microcode loads) , the LDA bus 91 (which provides decode addresses from the IPP 8, the jump address signals from JA bus 111 (which provide conditional sequencing information from EXl 10), the . output bus 112 from RCH 106 )which contains the present _,15 micro-address) , and bus 113 from the output of the microcode stack 107. This stack 107 holds addresses which are used to return from a microcode subroutine or from a microcode fault or exception sequence. The stack 107 can contain up to 16 addresses at once in

20 order to handle cases such as subroutine calls within subroutines. The 72-bit control output bus 110 of the microstore 104 is driven by way of buffers 105 over the RCC bus 64 to units 8-11 to provide microcode control of those units.

25 Execution 1 and Execution 2

The execution unit of the present embodiment performs the data manipulation and write-storage por- tions of all instructions which proceed through the dual pipeline (IP and EP) . Among the data types sup- 30 ported by this execution unit are:

1. 16 and 32-bit fixed point binary

2. 24-bit fraction/8-bit exponent floating point (single precision)

OMPI

3. 48-bit fraction/16-bit exponent floating point (double precision)

4. 96-bit fraction/16-bit exponent floating point (quad precision)

, 5. Varying length 8-bit character strings

6. Varying length 4 or 8-bit decimal digit strings

In the present embodiment the execution unit is located on two boards: EX 10 and EX2 11. The exe- cution unit operates under the control of microcode stored on the MCS 12. The microcode control bits are loaded into the RCM register 145 from bus 64. The exe¬ cution portion of a machine instruction may require one or many micro-instructions to complete. A new micro- instruction is fetched from the MCS 12 for each new data manipulation performed by EXl 10 and EX2 11.

The execution unit includes a general purpose 48-bit ALU 118 with an A-leg input and a B-leg input, selectors 117,125 for choosing among a plurality of operands for input to either the A- or B-leg, a selec¬ tor 121 for supporting operations on various data types, decimal and character string processing support networks 119,120,131, registers RS 126 and RD 122 for temporary data storage, a register file 130 and multiply hardware 133,146,147.

In the present embodiment, the ALU 118 is adapted to operate on data types up to 48 bits wide and provides a plurality of arithmetic and logical modes. Arithmetic modes include both binary and binary coded decimal types. The ALU 118 operates in concert with

- JH

OMPI ITU TE -ss..— » shift rotate network 119 and decimal network 122 to adaptively reconfigure in a manner permitting processing the various data types which must be pro¬ cessed.

A register file 130 supports separate read

(source) and write (destination) addresses for the instruction. The file 130 is 256 location deep and generally operates as a 32-bit wide file. In floating point arithmetic, field address register manipulation and certain other special cases, it supports a full 48- bit data path. An RF source decode 303 generates addresses for reading the register file 130 during the first phase of the OE stage while the RF destination decode 304 generates addresses for writing to the file 130 during the second phase of the ES stage. The RF destination decode 304 also transfers register update information to the collision detection logic 211 on the IPP 8 via bus BII 204. Selector 307 chooses between read and write addresses and sends those addresses to the register file 130.

The multiply hardware 133 consists of a 48- bit combination carry propagate/carry save adder. This adder 133 is combined with the sum register 146 and the carry register 147 to perform multiplications up to 48- by-48 bits by a shift and add technique. Each itera¬ tion of the multiply hardware 133 processes two bits of operand and generates two bits of sum and one bit of carry. The carry bit between the two sum bits is allowed to propagate.

Busses BBH 63 and BBL 62 supply to the execu¬ tion unit either a memory operand from the SPC 9 or a register or immediate operand from the IPP 8. This operand is latched in OPH 116 and OPL 123 which in turn

SUBS" TOTESHEET feed the B-leg selector 117 by way of busses 134 and 144 respectively. When the operand supplied over BBH 63 and BBL 62 is an unpacked 8-bit decimal digit data type, the decimal support logic 131 converts to the corresponding packed (4-bit) decimal data type.

The selector 117 selects from the destination register RD 122, OPH 116 and OPL 123 to drive the bus 135 which in turn feeds the B-leg of the main ALU 118. The A-leg selector 125 selects from among the input register RI 129 (which contains operands read from the register file 130), the shifter-register RS 126, the sum. bits bus 140 and carry bits bus 141 (output from the multiply hardware 133), the bus 132 (from the low word of the program counter RP 128), and the timer output 124 to drive the 48-bit A-leg ALU bus 143. The timers 124 are two general purpose counting registers used for operating system and performance evaluation support.

Program counter RP 128 is a 16-bit counter which can increment either by one or two depending on the length of the instruction currently in the execu¬ tion pipeline. If a jump or branch type of instruction is being processed, RP 128 may be loaded. This load occurs conditionally depending on whether the program is actually switching to a new non-sequential address and whether this change of flow was successfully pre¬ dicted by the branch cache 34 in the SPC 9. As described below, status about the branch cache's pre- dicton associated with the instruction currently in the execution unit is passed to EXl 10 by the IPP 8. In operation, the ALU 118 processes the data on busses 135 and 143 and the result is placed on bus 136. Bus 136 is coupled to the jump condition generation logic 300 which supplies microcode branching bits for loading into the JC REG 301. The contents of the JC REG 301 can effect the formation of the next microcode address

SUBST3TL?τcr SHEET either in the micro-instruction which loads it or in the one which immediately follows it. The control is effected by microcode control of the overlap of the OE stage of one instruction with the CF stage of the next one. Selector 302 chooses among a plurality of jump conditions to produce jump address signals which are transferred by way of JA bus 315 to the MCS 12.

Character byte rotation and floating point shifting are performed by the shift/rotate hardware of shift rotate network 119. Additional decimal digit processing, including unpack (convert 4-bit to 8-bit) and nybble rotate, is performed by network 120. The selector 121 chooses among its various sources depending on the data manipulation being performed. Selector 121 drives bus 137 which in turn loads RD 122, RS 126 and RP 128. This bus can also be coupled to busses BDH 50 and BDL 54 by the selector 127. The out¬ put bus 138 of RS 126 is selected onto BDL bus 50 and BDL bus 54 by the selector 127 in order to provide update information to the IPP 8 when an instruction completes execution which has modified a register which has a copy in the IPP 8. The output of RS 126 is also used to provide write data for the register file 130, to provide one of the operands to the multiply hardware 133 and as an input to the selector-125.

As described fully below, the use of RS 126 as an input to selector 125 is primarily for register bypassing. The register bypass logic 305 compares the register file source address (from source decode 303) for the instruction in stage OE to the register file destination address (from destinction decode 304) for the instruction in stage ES of the execution pipeline. If a match is detected, the contents of RS 126 on bus 138, which contains the data to be written into the register file 130 are selected by 125 (in place of the data read into RI 124 from the register file 130.)

BRANCH CACHE

The branch cache network is shown in Fig. 11. In the present embodiment, as shown in Fig. 11, por- tions of this network are located units 8-11. The branch cache network is adapted to permit predictions of non-sequestial program flow following a given instruction prior to a determination that the instruc¬ tion is capable of modifying instruction flow. Moreover, the branch cache network does not require computation of the branch address before the instruc¬ tion prefetching can continue. Generally, the branch cache network makes predictions based solely on the previous instruction locations, thereby avoiding the wait for decode of the current instruction before pro¬ ceeding with prefetch of the next instruction. Thus, the branch address need not be claculated before pre¬ fetching can proceed, since target addresses are stored along with predictions.

Detailed Explanation of Branch Cache Operation

In operation, the network shown in Fig. 3 begins on the SPC 9 with IRPL 33 accessing the branch cache 34 with same value that is being used to access thirty-two bits of instruction data in the program cache hardware 40,41,42,43. The output of the branch cache 34 includes a prediction bit (TAKEBR)_. (associated with the last word of a particular branch instruction and which asserts that a branch should be taken) , an index (which ensures the entry belongs to the current value of IRPL 33, a 16-bit target address (which will be loaded into IRPL 33 if the control indicates that non-sequential program flow should be followed), and a control line (ODDSIDE) (which indicates which of the two words of instruction data being fetched from the cache 41 a branch directive is associated with) . The signal ODDSIDE identifies each entry in the branch cache as being associated with either an odd or even wprd aligned instruction. In cases where a prediction is made for a two word branch instruction, the predic¬ tion entry is always associated with the second word of the instruction in order to ensure that the second word (which is required for calculating the address spe¬ cified by the branch instruction) is-properly fetched into the pipeline. The index and upper bits 1-7 are checked for equality in a comparator 218. If these values match and the signal TAKEBR indicates that the branch should be taken, the signal BCHIT is generated, causing the 16 bit target address (BTARG1-16) to be loaded into IRPL 33 via selector 32, rather than the normal operation of incrementing IRPL 33. The SPC 9 always sends the contents of the low side of the look- ahead program counter to the IPP 8 through buffer 35 where it is saved in register 217 for later use in validating the prediction. Many conditional instruc¬ tions in the Prime Instruction Set have branch addresses that are capable of being variable. For example, a conditional instruction could specify a branch to RP +• X, where RP = the contents of the program counter and X = the value of the index register. Between the time the branch cache was loaded with a target for a branch instruction and the time the instruction is actually executed, the value of the X register could change. In view of this possibility, the IPP 8 compares branch targets used for prefetching in the SPP 9 against the actual calculation of the location that the instruction will branch to if the specified conditions are satisfied. The calculation of

SUBSTITUTESHgET the address to which a branch instruction will vector is performed in the same manner as the generation of an address for a data operand. Therefore, the calculation performed in the AG stage of the IP produces the address to which the branch instruction should vector if the specified conditions are met. This address is eyentually passed to the EP for use in loading the program counter RP on EX2 11, and for use in reloading IRPL 33 on the IPP 8 if prefetching has not occurred properly, i.e. the branch cache makes an incorrect pre¬ diction. The calculated target is available on bus 90 from the last ALU 76 used in the AG stage. The calcu¬ lated target is compared to the value of the program counter (saved in REG 217), which contains the target prediction from the branch cache that was used to fetch the instruction following the branch instruction. Comparator 219 performs the equality check and indica¬ tes whether or not the. computed target address of the next instruction matches the target retrieved from the branch cache 34. If the equality is met, the signal GOODBRTARG is generated. Control logic 220 receives instruction classification information from decode net 82 and the BCHIT signal from the SPC 9 and determines whether or not a branch has occurred on a non-branch instruction. If such a branch has occurred, logic 220 generates the signal BREXCPTN. In another case, logic 220 and synchronizes the BCHIT signal from the SPC 9, passing it along with its associated instruction as BRTAKEN.

The signals GOODBRTARG, BRTAKEN, BREXCPTN are transferred to the branch processing hardware 221 in EXl 10 as the branch instruction enters the OE stage. As the branch instruction is executed, a determination of whether or not the branch should occur is loaded into register JCR 301. The output of register JCR 301

SUBSTITUTE SHEET OMPI al¬ together with GOODBRTARG and BRTAKEN are used to generate FLDRP which is used to force a load of RP 128 in EX 2 11 in the event the branch cache mechanism correctly predicted that a branch should be taken.

If the instruction flow has been correctly predicted, regardless of the outcome of the branch instruction, the signal CEXCMPL, indicating that no further execution cycles are required in the EP, is to the PCU 1, which allows the IP to proceed.

In the event that the branch cache mechanism has not correctly predicted program flow, further exe¬ cution cycles in the EP are necessary. Bus JA 315 transfers the address of the next microstep (from JCR 301) thereby specifying which type of branch cache modification is to be performed.

Modifications may be one of two categories for branch-type instructions, depending on the probabi¬ lity of correct prediction of branches. For both pre¬ dictable and non-predictable instructions, if the instruction is incorrectly predicted to branch, the branch cache 34 is updated by removing the prediction while permitting the "bad" target address to remain.

If a branch occurs which has not been pre¬ dicted on an instruction type which is classified as "predictable" (such as a Jump or Branch instruction), the branch cache 34 is updated during the ensuing exe¬ cution cycles by inserting a prediction and associated target address. The newly inserted target address, which is the calculated address of the branch instruc- tion, is transferred from selector 127 by way of BDL bus 54 to branch cache 34. Where the branch is ' correctly predicted, but the target address does not match the calculated target address, the prediction remains in the branch cache 34 but a new target address (corresponding to the calculated address) is inserted.

If a branch occurs which has not been pre- dieted for instruction types which are not classified as "predictable" (such as Skip) , no updating is made in the branch cache 34.

Where a branch is incorrectly predicted for an instruction which is not a branch-type instruction, the signal BREXCPTN forces execution of a microcode ^* routine not associated with any particular instruction which removes the incorrect prediction category. In all cases of an incorrect prediction, the look-ahead program counter IRPL 33 is reloaded and the PCU 1 is notified to flush the pipeline.

REGISTER BYPASS

The register bypass network is shown in detailed form in Fig. 12. In the present embodiment, the register bypass network is located principally on IPP 8. In the present pipelines system, simultaneous access to certain registers is often required by two or more different stages of the pipelines. For example, many instructions require prefetching of certain registers early in the pipeline sequence so that they may be used in the generation of data (operand) addresses for accessing the program storage. Other instructions require prefetching of a register value which is used directly as an operand. Register values used for generating addresses, or directly as operands are typically modified by execution stages placed late in the pipeline.

SUfaiS s i 1! ' ' « ϊ_SΪ^*iϊ__---ⁱi With this type of processor, instruction "collisions" may occur when two instructions, one pre¬ fetching a register and one writing it, are too close to each other in the instruction flow. In this situation, the write which happens in a late stage may not actually be done until later in time than the pre¬ fetch read, even though the writing instruction comes before the reading one in the program.

The register bypass network accommodates hardware which handles collisions between an instruc¬ tion reading a register in an operand prefetch stage of a pipeline and another instruction modifying the same register in an execution stage which may be employed to modify many registers during one instruction through repeated execution cycles. The register bypass network further accommodates different types of collision using variations of bypassing techniques. If collision occur on instructions which are well separated, a bypass selector and associated storage for saving the bypass value are sufficient, together with address comparison hardware. As the two instructions move closer together and the prefetched register is being used to form an operand address, the pipeline control unit PCU 1 forces separation of the instructions; however, this separation only occurs if a collision is either detected or at least predicted. The register bypass further provides routing bypass data back to different stages of the pipeline depending on the relative separation in cases where register prefetching is only occurring on behalf of register operands rather than register-related operand address formation. In the register bypass net¬ work of Fig. 12, a pair of registers are fetched for each memory referencing instruction. These registers are termed "base register" and "index register", and are shown as AGRF 72 in Fig. 8. The base and index

SUBST-T'.. EET register are added together by ALU 75 in the AG stage of the instruction fetch pipeline, thence added to a displacement resulting in an operand address.

Another instruction form requires that the value of a "general register" be supplied directly as an operand. This operand is fetched from the same register file as is used for the base registers described above, and is transported without modifica¬ tion through the AG stage and supplied to the OE stage.

Current values for base, index, and general registers are supplied by the ES stage as it executes microcode instructions which modify them. The ES stage can modify all 32 bits of a register, or either of its 16 bit halves. Since the ES stage completes its opera- tions three stage times later than completion of the corresponding ID stage, there are three different collisions possible:

1) Modification and use separated by three or more cycle times. In this case, an instruction has completed the ID phase and waits for completion of the terminal microcode step of the preceding instruc¬ tion before continuing through the AG phase. An index and base register have been fetched from the AG Register file

72, transferred through pipeline register BXR 73 and stored in selector/latch 74. The register file destination address specified by each microcode step and supplied by BII 204 is continuously compared (by comparators 226 and 227) with the base register and index register addresses used on behalf

SUBSTITUTESHEET

^{~ "} of the instruction awaiting in the AG stage and stored in latch .225. The out¬ puts of these comparators, together with write enables supplied by BII, are 5 passed through bypass control logic 228 for determination of the needed action.

If a match occurs, the data in selector/ latch 74 is stale, and correct data must be substituted. The appropriate por-

10 tions of selector/latch 74 are re- clocked, selecting the updated value coming from the ES stage via BD 50, 54, buffers 210 and pipeline register BDR 71. Sufficient time exists in this

15 case for the updated values to re- traverse the AG stage, so no additional delay is necessary. This same mechanism is employed for equivalent cases involving general registers used as

20 operands.

2) Modification and use separated by two cycle times.

In this case the AG phase is attempting to proceed (the final microcode step of

25 the preceding instruction is beginning) and the previous microcode step modified an index or base register used by the instruction active in the AG phase. The same monitoring hardware used for 1

30 remains effective due to latch 225, which holds the index and base register addresses long enough for this final determination. In the event of colli-

OMPI sion detection, the proper bypass is again selected at selector/latch 74, but in this case extra time must be added for the AG phase to properly employ the

5 new value. The Collision Detect signal, produced by control logic 228, directs the PCU to allow the ES stage to complete while stopping all other pipe¬ line stages. In this fashion the new

10 value is obtained and a one cycle time delay provided for the AG phase to make use of it.

It is undesirable to incur this time delay where registers are used directly

15 as operands. Since this type of operand need not be manipulated by ALU's 75 and 76, it is possible to skip over these pipeline stages and send the data directly where it's needed. This is

20 accomplished via selectors 212 and 213, which select the modified portion of the value presently on busses BD 50, 54 for insertion into the data stream in place of the stale value being produced on

25 busses 89 and 90. In this manner, no extra time is required.

3) Modification and use separated by one cycle time.

When two successive machine instructions 30 result in this situation, the method used in 1) and 2) is not effective, because the instruction with the stale data must exit the AG stage before the

SUEc_" - r--J _ S5U ST OlvSPI

, register file destination address of the modifying instruction is available. The destination predictor logic, consisting of a portion of the decode net 82, cer-

5 tain saved opcode bits 207 and control logic 229, is used to determine which » register, if any, might be modified in the final microcode step of an instruc¬ tion. This requires some care in the

10 selection of microcode algorithms, but the flexibility resulting from storage of control bits in the decode net makes this task straightforward.

The output of the destination predictor 15 logic is compared with the index and base register addresses used by the next instruction by comparators 230 and 231. The outputs of the comparators travel through control logic 233, which genera- 20 tes the Collision Predict signal. When asserted, this signal instructs the PCU to allow the instruction doing the modi¬ fication to proceed, while holding the next instruction's AG stage (and all 25 subsequent instructions) . This separa¬ tes the two instructions by two cycles instead of one cycle, and the hardware of case 2) above can then take over. This logic may or may not insert 30 its one cycle delay, depending on whether the collision actually occurs.

It is again undesirable to apply time penalities when registers are used as operands. When a match is detected by

comparator 230 and a general register is being fetched, this condition is remem¬ bered in register 232. This is in turn pipelined in Register 234 and sent over to the OE stage hardware as the signal

USERD, where it acts as a form of , extended control over the operand source select microcode field. When such a collision occurs, this extended control forces selection of the needed operand from an alternative source in the instruction execution pipeline. This extra copy is kept valid by microcode convention, and again no time penalty is required.

The invention may be embodied in other speci¬ fic forms without departing from the spirit or essen¬ tial characteristics thereof. The present^* embodiments are therefore to be considered in all respects as illustrative and not restrictive, the scope of the invention being indicated by the appended claims rather than by the foregoing description, and all changes which come within the meaning and range of equivalency of the claims are therefore intended to be embraced therein.

GU?I

Claims

What is claimed is:

1. A data processing system for processing a sequence of program instructions comprising an instruction pipeline having a plura- lity of serially operating instruction stages for reading instructions from storage and for forming therefrom data to be employed during execution of said instructions, an execution pipeline having a plurality of serially operating execution stages for receiving said data and for employing said data formed by said instruction pipeline for executing said instructions, a pipeline control unit for synchronously operating said instruction pipeline and said execution pipeline, said pipeline control unit including means for initiating operation of at least one stage of said execution pipeline using data formed by said instruction pipeline for a program instruction prior to the completion of said data for- mation by said instruction pipeline for said program instruction, whereby operation of at least one instruction stage and one execution stage of said respective pipelines overlaps for each program instruc- tion.

2. The data processing system of claim 1 further comprising a high speed random access memory, a pipeline master clock for timing said pipeline stages, said pipeline control unit providing at least two clocked periods for each said pipeline stage to complete its operation, and for each said at least two clocked

SUBSTΓΓU^" . Eit;ET periods, said instruction pipeline having access to said high speed memory during one of said clocked periods and said execution pipeline having access to said high speed memory during another one of said clocked periods.

,

3. The data processing system of claim 1 wherein said pipeline control unit further comprises means responsive to exception conditions on said execution and said instruction pipelines for independently controlling, for each pipeline, the flow of instruction operations through said execution pipe¬ line and said instruction pipeline.

4. The data processing system of claim 3 wherein said flow control means includes means for halting operation of one only of said execution and instruction pipelines.

5. The data processing system of claim 1 wherein said instruction pipeline comprises an instruction fetch stage for accessing from memory program instructions to be performed, an instruction decode stage for generating, from said accessed instructions, (a) starting addresses in a microcode storage element and (b) operand address data, and an address generation stage for generating operand addresses from said operand address data.

6. The data processing system of claim 5 further comprising a register file and wherein said execution pipeline comprises a control formation stage for accessing microinstructions from said microcode storage element.

SUESTJTUTΞSHEET using said starting addresses, and for buffering said microinstructions, an operand execute stage for accessing, using said operand addresses and register file controls, operand data to be operated upon and, using said operands and a said microinstruction, initiating execution of said instruction, and an execution and store stage for completing said execution of said microinstruction and making results of said execution available to said system.

7. The data processing system of claim 6 wherein said operation initiating means begins opera¬ tion of said control formation stage for a program instruction at a time prior to completion of operation of said address generation stage for said program instruction.

8. The data processing system of claim 6 wherein the pipeline control unit further comprises means responsive to said microinstruc¬ tions for altering the flow of instructions in at least one of said instruction pipeline and said execution pipeline.

9. The data processing system of claim 8 wherein said altering means is responsive to a said microinstruction for extending the operating time dura¬ tion of all stages, except the execute and store stage, for allowing the operand execute stage to complete a process operation.

10• The data processing system of claim 8 wherein said altering means is responsive to a said microinstruction for inhibiting operation of said

instruction pipeline for allowing said execution pipe¬ line to cycle through a plurality of microinstructions.

11. The data processing system of claim 8 wherein said pipeline control unit further comprises means for inserting no-operation cycles into the execution pipeline in the event that jat least one of (a) being no instruction from the instruction decode stage of the instruction pipeline for the control formation stage of the execution pipeline and (b) there being no instruction from the address genera¬ tion stage of the instruction pipeline for the operand execution stage of the execution pipeline, occurs.

12. The data processing system of claim 8 wherein said altering means is responsive to a con- ditional branch microinstruction entering said operand execute stage and said altering means further comprises means for operating said operand execute and execution and store stages, and means for inhibiting operation of said instruction pipeline and said control formation stage, until data required by said conditional branch micro¬ instruction is available from said operand execute stage.

13. The data processing system of claim 6 wherein said instruction pipeline comprises a look-ahead program counter, and further wherein said pipeline control unit comprises means responsive to a said microinstruc- tion for redirecting instruction flow in said instruc¬ tion pipeline by effecting reloading of said instruction pipeline look-ahead program counter.

OMPI

14. The data processing system of claim 13 wherein said execution pipeline includes a microcode storage element, and further wherein said pipeline control unit, in response to a request by the execution pipe¬ line causes all current instructions in the instruction pipeline to be discarded.

15. The data processing system of claim 1 wherein said execution pipeline includes a microcode storage element and further wherein said pipeline control unit, in response to a request by the execution pipe¬ line will discard all current instructions in the instruction pipeline.

16. The data processing system of claim 15 wherein said pipeline control unit further comprises means for inserting no-operation cycles into the execution pipeline during the time duration that the instruction pipeline is refilling, and for continuing operation of said execution pipeline while said instruction pipeline is refilling.

17. The data processing system of claim 1 comprising means for detecting pipeline collisions in said instruction and execution pipelines, and further wherein the pipeline control unit comprises means responsive to said collision detecting means for delaying operation of at least one of said stages for introducing a separation between said colliding instructions.

18. The data processing system of claim 5 comprising eans for detecting an exception con¬ dition during operation of said instruction fetch stage and wherein said pipeline control unit comprises means for hold an instruction in said instruction fetch stage until all other stages of said instruction and execution pipelines have completed pro¬ cessing instructions therein.

19. The data processing system of claim 5 further comprising a high speed instruction storage element, means for reading from said element two instruction words at a time, said two words being aligned with an even word boundary of said memory, and access means for reading from said ele¬ ment a two word instruction aligned with an odd word boundary said access means comprising means for reading- a first word of said two word instruction during the instruction fetch stage of said instruction, said first work including all instruction decode data, and means for reading a second word of said instruction during the instruction fetch- stage of a next following instruction.

20. The data processing system of claim 1 further comprising a microcode storage element for storing microinstructions, and said execution pipeline effects data manipulation in response to selected ones of the microinstructions.

5--- _ S _ *J^* _ __ ■-*>- _-____. t

OMPI

21. In a data processing system for pro¬ cessing a sequence of program instructions comprising an instruction pipeline having a plura¬ lity of serially operating instruction stages for 5 reading instructions from storage and for forming therefrom address data to be employed during execution of said instructions, an execution pipeline having a plurality of serially operating execution stages for receiving 10 said address data and for employing said address data formed by said instruction pipeline for referencing stored data to be employed for executing said instruc¬ tions, the pipeline control method comprising ,15 steps of synchronously operating said instruc¬ tion pipeline and said execution pipeline, and initiating operation of at least one stage of said execution pipeline using at least one 20 said address data formed by said instruction pipeline for a program instruction prior to the completion of said address data formation by said instruction pipe¬ line for said instruction.

22. The pipeline control method of claim 21 25 further comprising the steps of providing at least two clocked periods for each pipeline stage to complete its operation, and sharing a high speed memory between said instruction pipeline and said execution pipeline, said 30 instruction pipeline having access to said high speed memory during one of said clocked periods and said execution pipeline having access to said high speed memory during another one of said clocked periods.

23. The pipeline control method of claim 21 35 further comprising the step of

luώo - _ _ . __ S-s shT _/^ξuRE

_ OMPI independently controlling, for each pipe¬ line, the flow of instruction operations through said respective execution and instruction pipelines.

24. The pipeline control method of claim 23 wherein said controlling step further comprises the step of halting operation of one only of said execution and instruction pipelines in response to pipeline control conditions.

25. The pipeline control method of claim 21 further comprising the steps of detecting pipeline collisions in said instruction and execution pipelines, and delaying operation of at least a portion of said instruction pipeline for introducing a separa- tion between said colliding instructions.

26. A data processing system for processing a sequence of program instructions comprising an instruction pipeline having a plura¬ lity of serially operating instruction stages for reading instructions from storage and for forming therefrom plural address data to be employed during execution of said instructions, an execution pipeline having a plurality of serially operating execution stages for receiving said address data and for employing said address data formed by said instruction pipeline for referencing stored data to be employed for executing said instruc¬ tions, a pipeline control unit for operating said instruction pipeline and said execution pipeline, said pipeline control unit including means responsive to exception conditions on said execution and said instruction pipelines for

independently controlling, for each pipeline, the flow of instruction operations through said execution pipe- and said instruction pipeline.

27. A data processing system for processing a sequence of program instructions comprising

. an instruction pipeline having a plura¬ lity of serially operating instruction stages for reading instructions from storage and for forming therefrom plural address data to be employed during execution of said instructions, an execution pipeline having a plurality of serially operating execution stages for receiving said address data and for employing said address data formed by said instruction pipeline for referencing stored data to be employed for executing said instruc¬ tions, and said execution pipeline having a recon- figurable arithmentic logic unit responsive to control signals generated in accordance with said program instructions.

28. A data processing system for processing a sequence of program instructions comprising an instruction pipeline having a plura¬ lity of serially operating instruction stages for reading instructions from storage and for forming therefrom plural address data to be employed during execution of said instructions, an execution pipeline having a plurality of serially operating execution stages for receiving said address data and for employing said address data formed by said instruction pipeline for referencing stored data to be employed for executing said instruc¬ tions, a pipeline control unit for operating

_ _ "p—;^-.- ,: „ _. - - - ' * » ., , -. e --_* « / QJViPI

'"-=--=."1 "J p _ ^.- ^. . . said instruction pipeline and said execution pipeline, a microcode storage element for storing microinstructions accessed in accordance with said program instructions, o an arithmetic logic unit, and means operable with the arithmetic logic unit for reconfiguring said arithmetic unit for per¬ mitting processing of data in alternate format presen- . tation.

29. The data processing system of claim 28 wherein said reconfiguring means comprises a shift and rotate circuitry, and a decimal conversion circuitry.

30. A program instruction flow prediction apparatus for a data processing system having means for prefetching an instruction, said flow prediction apparatus comprising a flow prediction storage element, an instruction storage element containing program instructions to be executed, means for addressing an instruction in said instruction storage element, means for reading a flow control word in said flow prediction storage element at a location derived from the location of said addressed instruction in said instruction storage element, said flow control word containing at least a branch control portion for predicting the flow of program instructions and a next program instruction address portion containing at least a portion of a next program address if a program branch is predicted.

31. The program instruction flow prediction apparatus of claim 30 further comprising means for

^•■^TfflX

SUBSTITUTE SKΠT O PI monitoring instruction flow during said instruction prefetching and means responsive to said monitoring means for updating said flow prediction storage element based solely upon a history of the instruction flow.

32. The program flow instruction prediction apparatus of claim 31 wherein said flow prediction storage elements is a random access, high speed memory.

33. The program flow prediction apparatus of claim 32 further wherein each said flow prediction memory has significantly less storage capacity then a system main memory.

34. The program flow prediction apparatus of claim 31 further wherein said flow prediction storage element has fewer storage locations than said program instruction storage element.

35. The program instruction flow prediction apparatus of claim 31 wherein said monitoring means responds only to the instruction flow of a most recent execution of said present instruction.

36. The program instruction flow prediction apparatus of claim 30 further comprising means for employing a flow altering data from said flow prediction storage element in place of a next sequential flow data during a next program^* decode operation.

37. The program instruction flow prediction apparatus of claim 31 further wherein said monitoring and update means comprise means for updating said prediction

OMPI storage element only when actual program instruction flow does match the prediction data stored in said pre¬ diction storage element.

38. The program instruction flow prediction apparatus of claim 30 wherein said instruction address portion contains a portion of a next program address and further comprising means for combining said address portion with said instruction memory location address for forming the address location of a next instruction to be fetched.

39. The program instruction flow prediction apparatus of claim 34 wherein said flow control word further includes an index portion, and further comprising means for comparing said index portion with the index associated with a current instruction for inhibiting false prediction of a branch condition for non-branch instructions which map into the same location of said flow prediction storage element as did valid branch instructions.

40. The program instructions flow prediction apparatus of claim 38 wheein said data processing system has an instruction and execution pipeline means having a plurality of serially operating stages, one of said stages being an instruction fetch stage having a look-ahead program counter, and further comprising means for loading said counter with flow altering data flowing said flow prediction storage ele¬ ment during a normal operation time duration of said fetch stage.

S 3 J J^ Λ*-.,

SO _ _ l r

HEET OMPI

41. The program instruction flow prediction apparatus of claim 37 further wherein said updating means operates to remove a false branch control portion data.

42. A program instruction flow prediction method for a data processing system including the step of prefetching instructions, said flow prediction method comprising the steps of addressing an instruction in an instruc- tion storage element, addressing a flow control word from a flow prediction storage element at a location derived from the location of said addressed instruction in said instruction storage element, said flow control word containing at least a branch control portion for pre¬ dicting the flow of program instructions and a next program instruction address portion containing at least a portion of a next program address if a program branch is predicted, and deriving from the flow control word a new next program address if a branch is predicted.

43. The program instruction flow prediction method of claim 42 further comprising the steps of monitoring instruction flow during said instruction prefetching, and updating said flow prediction storage element based solely a history of a the instruction flow.

44. The program instruction flow prediction method of claim 42 further comprising the step of employing a flow altering data from said flow predic¬ tion storage element prior in place of a next sequen¬ tial flow data during a next program instruction fetch.

-TT

45. The program instruction flow prediction method of claim 43 further wherein said updating step comprises the step of updating said prediction storage element only when actual program instruction flow does not match the prediction data stored in said prediction storage element.

46. The program instruction flow prediction method of claim 42 further comprising the step of com¬ bining said instruction address portion with said instruction memory location address for forming the address location of a next instruction to be fetched.

47. The program instruction flow prediction method of claim 43 further comprising the steps of storing as part of said flow control word an index portion, comparing said index portion with the index associated with a current instruction, and inhibiting false prediction of a branch condition for a nonbranch instruction which maps into the same location of said flow prediction storage ele¬ ment as the valid branch instructions as a result of said index comparison being invalid.

48. A data processing system for processing a sequence of program instructions comprising an instruction pipeline having a plura¬ lity of serially operating instruction stages for reading instructions from storage and for forming therefrom plural address data to be employed during execution of said instructions, an execution pipeline having a plurality of serially operating execution stages for receiving said address data and for employing said address data formed by said instruction pipeline for referencing

__ OMPI stored data to be employed for executing said instruc¬ tions, a pipeline control unit for synchronously operating said instruction pipeline and said execution pipeline, said pipeline control unit including a plurality of state registers, a plurality of combinatorial logic cir¬ cuits, one each of said state registers and said logic circuits being associated with each stage of said pipelines, each said logic circuit having a first signal output and a second signal output, each pipeline stage having a first phase of operation associated with said first signal output of said associated logic circuit and a second phase of operation associated with said second signal output of said associated logic circuit, each said logic circuit and associated state register, associated with the same pipeline, being connected in series, and at least one of said logic circuits being connected to receive condition signal from said pipeli- nes for controlling the flow of instructions through said pipeline.

49. The data processing system of claim 48 further comprising means for connecting said first signal output of a logic cirucit to the associated state register for determining when the associated pipeline stage has completed a first phase of operation.

50. A data processing system for processing a sequence of program instructions comprising an instruction pipeline having a plura¬ lity of serially operating instruction stages for reading instructions from storage and for forming therefrom plural address data to be employed during execution of said instructions, an execution pipeline having a plurality of serially operating execution stages for receiving said address data and for employing said address data formed by said instruction pipeline for referencing stored data to be employed for executing said instruc¬ tions, a pipeline control unit for synchronously operating said instruction pipeline and said execution pipeline, said pipeline control unit including means for initiating operation of at least one stage of said execution pipeline using one said address data formed by said instruction pipeline for a program instruction prior to the completion of said address data formation by said instruction pipe- line for said program instruction, whereby operation of at least one instruction stage and one execution stage of said respective pipelines overlaps for each program instruc¬ tion.

51. A data processing system for processing a sequence of program instructions comprising an instruction pipeline having a plura¬ lity of serially operating instruction stages for reading instructions from storage and for forming therefrom plural address data to be employed during execution of said instructions, an execution pipeline having a plurality of serially operating execution stages for receiving said address data and for employing said address data formed by said instruction pipeline for referencing

SUESTΓΠJTΞ SHEET stored data to be employed for executing said instruc¬ tions, a pipeline control unit for operating said instruction pipeline and said execution pipeline, means for detecting collisions between data read from a register in the instruction pipe¬ line phase of operation in response to a first instruc¬ tion and write data written in registers during the execution phase of operation in response to an earlier instruction, wherein said execution phase can include a plurality of execution cycles during each of which a register can be modified and wherein said first instruction requires one of said modified values to continue valid operation, said detecting means comprising, means for storing said modified values generated during the execution phase and the write register address associated therewith, means for comparing the associated write register address of each modified value with the read register address read by the instruction pipeline,

.means for directing, when said addresses match, the modified value, to be written at said register address to replace the data previously designated to be used during said instruction phase of operation.

52. The data processing system of claim 51 wherein said directing means further comprises a selector responsive to said comparing means for selecting the data read during the instruc¬ tion phase or said modified value, and a storage register for storing the last written modified value generated during said execution phase of operation.

53. The data processing system of claim 52 further comprising a microcode storage element, and means for reading microinstructions from said storage element for controlling operation of said detecting means.

54. The data processing system of claim 51 wherein said directing means further comprises means for detecting collisions between a portion of the modified register and a corresponding portion of the read data, and said directing means replaces only said portion of the read data with the portion of the modified data value.

55. The data processing system of claim 54 wherein said register portion is one-half of the read data.

56. The data processing system of claim 51 further wherein said detecting means comprises means for generating a pipeline collision signal in response to a detected collision, and means for generating an execution stage signal indicating that generation of the modified data has not been completed, and wherein said pipeline control unit comprises means responsive to said pipeline colli¬ sion signal and said execution stage signal for inhi¬ biting operation of all pipeline stages other than those required to generate said modified data until said modified data has been generated.

57. The data processing system of claim 51 wherein said directing means includes

means for directing said modified data to a plurality of pipeline stages, each said stage then using said modified data in place of said originally read data.

58. The data processing system of claim 6 further comprising means for detecting collisions between read data from a register associated with the instruc¬ tion pipeline phase of operation in response to a first instruction and write data written in registers asso¬ ciated with the execution pipeline phase of operation in response to an earlier instruction wherein said exe¬ cution phase of operation can include a plurality of execution cycles during each of which a register can be modified and wherein said first instruction requires one of said modified values to continue valid opera¬ tion, said detecting means comprising means for storing said modified values generated during the execution phase and the write register address associated therewith, means for comparing the associated write register address of each modified value with the read register address used by the instruction pipeline, means for directing, when said addresses match the modified value, to be written at said register address, to replace the data previously designated to be used during said instruction phase of operation.

59. The data processing system of claim 58 wherein said storing means receives data from the execution and store stage, said read register address is a read

SUL ^t JTUTΞ SHEET address generated by the instruction decode stage, said write register address is available during operation of the execute and store stage, and. said directing means comprises a selector means connected between the instruction decode and address generation stages, said selector having the read data and the modified data as inputs thereto.

60. The data processing system of claim 58 wherein said directing means further comprises a second selector means connected between the address generation stage and the operand execute stage for altering at least a portion of the flow of address data to said operand execute stage in response to a collision detection signal.

61. A data processing system for processing a sequence of program instructions, said system comprising an instruction pipeline having a plura¬ lity of serially operating instruction stages for reading instructions from storage and for forming therefrom plural address data to be employed during execution of said instructions, an execution pipeline having a plurality of serially operating execution stages for receiving said address data and for employing said address data formed by said instruction pipeline for referencing stored data to be employed for executing said instruc¬ tions, a pipeline control unit for operating said instruction pipeline and said execution pipeline, means for predicting collisions between read data from a register in the instruction pipeline phase of operation in response to a first program instruction and write data written in registers during the execution phase of operation in response to an earlier program instruction wherein the execution phase can include a plurality of execution cycles during each of which a register can be modified, and wherein said first instruction requires one of said modified values to continue valid operation, said prediction means comprising means for predicting a destination address for the earlier of two instructions, means for comparing the predicted destination address with an actual read address of a later instruction and for generating a predicted colli¬ sion signal when the two addresses match, and said pipeline control unit having means for inserting a delay between the two potentially colliding instructions wherein the write address can be formed before the read address data is used to access data.

62. In a data processing system for pro- cessing a sequence of program instructions, the data processing system having an instruction pipeline having a plura¬ lity of serially operating instruction stages for reading instructions from storage and for forming therefrom plural address data to be employed during execution of the instructions, an execution pipeline having a plurality of serially operating execution stages for receiving the address data and for employing the address data formed by the instruction pipeline for referencing stored data to be employed for executing the instruc¬ tions, a collision detection method comprising the steps of detecting collisions between read

_ )5U^'j^'ι5_-^i}«w^".T i 'J.^*T i 2i. data from a register during the instruction pipeline phase of operation and in response to a first instruc¬ tion, and write data to be written in registers during the execution phase of operation, and in response to an earlier instruction, wherein the execution phase can include a plurality of execution cycles during each of which a register can be modified, and wherein the first instruction requires one of the modified values to con¬ tinue valid operation, said detecting step further comprising storing the modified values generated during the execution phase and the write register addresses associated therewith, comparing the associated write register address of each modified value with the read register address read by the instruction pipeline, and selecting, when the addresses match, the. modified value, to be written into the matching write address, to replace the data previously designated to be used during the instruction phase of operation.

63. The collision detection method of claim 62 further comprising the steps of generating a pipeline collision signal in response to a detected collision, generating an execu¬ tion stage signal indicating that generation of the modified data has not been completed, and inhibiting operation of all pipeline sta¬ ges other than those required to generate said modified data until said modified data generation has been completed.

64. The collision detection method of claim 62 further comprising the step of directing the modified data to a plura-

SUBSTITUTΞ SHEET lity of pipeline stages, each stage using the modified data in place of the originally read data during the instruction phase of operation.

65. In a data processing system for pro- cessing a sequence of program instructions, the data processing system comprising an instruction pipeline having a plura¬ lity of serially operating instruction stages for reading instructions from storage, and for forming therefrom plural address data to be employed during execution of the instructions, an execution pipeline having a plurality of serially operating execution^" stages for receiving the address data and for employing the address data formed by the instruction pipeline for referencing stored data to be employed for executing said instruc¬ tions, a collision prediction method comprising the steps of predicting the destination address to be written by the earlier of two instructions which may potentially collide, comparing the predicted destination ₍. address with an actual read address of a later occurring instruction, and delaying processing of said later instruction for a time so that the address to be used for writing of said earlier instruction can be formed whereby a collision detection can be determined between said two instructions.

^{"" ***■}"" " O