Electrical computers and digital processing systems: memory – Storage accessing and control – Access timing
Reexamination Certificate
2002-03-14
2004-10-12
Sparks, Donald (Department: 2187)
Electrical computers and digital processing systems: memory
Storage accessing and control
Access timing
C711S150000, C711S151000, C711S167000, C711S210000, C709S241000, C709S241000, C712S206000, C712S215000, C712S216000, C712S219000, C712S246000
Reexamination Certificate
active
06804759
ABSTRACT:
FIELD OF THE INVENTION
The present invention relates generally to digital data processing, and more particularly to pipelined operations in a processing unit of a data processing system.
BACKGROUND OF THE INVENTION
A modern computer system typically comprises a central processing unit (CPU) and supporting hardware necessary to store, retrieve and transfer information, such as communications busses and memory. It also includes hardware necessary to communicate with the outside world, such as input/output controllers or storage controllers, and devices attached thereto such as keyboards, monitors, tape drives, disk drives, communication lines coupled to a network, etc. The CPU is the heart of the system. It executes the instructions which comprise a computer program and directs the operation of the other system components.
From the standpoint of the computer's hardware, most systems operate in fundamentally the same manner. Processors are capable of performing a limited set of very simple operations, such as arithmetic, logical comparisons, and movement of data from one location to another. But each operation is performed very quickly. Programs which direct a computer to perform massive numbers of these simple operations give the illusion that the computer is doing something sophisticated. What is perceived by the user as a new or improved capability of a computer system is made possible by performing essentially the same set of very simple operations, but doing it much faster. Therefore continuing improvements to computer systems require that these systems be made ever faster.
The overall speed of a computer system (also called the “throughput”) may be crudely measured as the number of operations performed per unit of time. Conceptually, the simplest of all possible improvements to system speed is to increase the clock speeds of the various components, and particularly the clock speed of the processor. E.g., if everything runs twice as fast but otherwise works in exactly the same manner, the system will perform a given task in half the time. Early computer processors, which were constructed from many discrete components, were susceptible to significant speed improvements by shrinking component size, reducing component number, and eventually, packaging the entire processor as an integrated circuit on a single chip. The reduced size made it possible to increase the clock speed of the processor, and accordingly increase system speed.
Despite the enormous improvement in speed obtained from integrated circuitry, the demand for ever faster computer systems has continued. Hardware designers have been able to obtain still further improvements in speed by greater integration (i.e., increasing the number of circuits packed onto a single chip), by further reducing the size of the circuits, and by various other techniques. However, designers can see that physical size reductions can not continue indefinitely, and there are limits to their ability to continue to increase clock speeds of processors. Attention has therefore been directed to other approaches for further improvements in overall speed of the computer system.
Without changing the clock speed, it is possible to improve system throughput by using multiple copies of certain components, and in particular, by using multiple CPUs. The modest cost of individual processors packaged on integrated circuit chips has made this practical. While there are certainly potential benefits to using multiple processors, additional architectural issues are introduced. Without delving deeply into these, it can still be observed that there are many reasons to improve the speed of the individual CPU, whether or not a system uses multiple CPUs or a single CPU. If the CPU clock speed is given, it is possible to further increase the speed of the individual CPU, i.e., the number of operations executed per second, by increasing the average number of operations executed per clock cycle.
Most modern processors employ some form of pipelining to increase the average number of operations executed per clock cycle, as well as one or more levels of cache memory to provide high-speed access to a subset of data in main memory. Pipelined instruction execution allows subsequent instructions to begin execution before previously issued instructions have finished. Ideally, a new instruction begins with each clock cycle, and subsequently moves through a pipeline stage with each cycle. Even though an instruction may take multiple cycles or pipeline stages to complete, if the pipeline is always full, the processor executes one instruction every cycle.
Of course, the pipeline being always full is simply an ideal towards which designers strive, knowing that it is impossible to always keep the pipeline full. For various reasons, the pipeline will sometimes stall. For example, the instruction stream may take an unexpected branch to an instruction which is not in the cache, or may load data from a data location which is not in the immediate (lowest level) cache. In these cases, the processor can not begin a new instruction, and must typically wait until the necessary instruction or data is fetched into the cache, either from another higher level cache, or from main memory.
There are other causes of pipeline stall. Among them are address conflicts between pipeline operations, particularly, between load and store operations. If a store operation stores data to an address X, and a load operation subsequently loads data from address X, care must be taken that the store operation completes before the load operation begins, or incorrect data may be loaded. In order to prevent erroneous operation, a processor using pipelined instruction execution typically compares the address of a target operand of certain operations with similar addresses of operations in the pipeline. If a conflict is detected, the subsequent operation must be delayed or restarted.
Some system designs, of which UNIX-based systems are an example, employ a form of virtual addressing which has the possibility of address aliasing. I.e., addresses derived from the instructions and generated by the processor, which are often referred to as “virtual addresses” or “effective addresses”, are mapped to addresses in the physical main memory of the system, generally referred to as “real addresses” or “physical addresses”, where it is possible that multiple virtual addresses map to the same real address. Because multiple virtual addresses may map to the same real address, the cache is typically accessed with a real address and only the real address may reliably be used to determine whether there is an address conflict in pipeline operations.
Typically, in order to obtain a real address of data for a data reference operation (e.g., a load or store operation), a portion of the virtual address is used to access a table called a translation lookaside buffer (TLB). The TLB is typically N-way set associative, providing N possible real address entries corresponding to the virtual address. A TLB lookup requires that the N entries be retrieved from the TLB, that each entry be compared to the virtual address, and that the real address corresponding to the matched entry be selected. These operations may require multiple clock cycles.
Where an address conflict exists between pipeline operations, it is desirable to detect the conflict as soon as possible. The longer it takes to detect such a conflict, the greater is the potential performance impact. Not only is the conflicting instruction potentially compromised, but instructions occurring after the conflicting instruction may be compromised as well. Late detection of an address conflict requires that all potential data integrity exposures be rectified before proceeding. Since an address conflict can not be detected until the virtual addresses are translated to real addresses, the time required to perform the TLB lookup directly delays the detection of an address conflict.
As processors grow more capable and more complex, the problem of address conflicts between pipeline operations will b
International Business Machines - Corporation
Sparks Donald
Truelson Roy W.
Truong Bao Q.
LandOfFree
Method and apparatus for detecting pipeline address conflict... does not yet have a rating. At this time, there are no reviews or comments for this patent.
If you have personal experience with Method and apparatus for detecting pipeline address conflict..., we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Method and apparatus for detecting pipeline address conflict... will most certainly appreciate the feedback.
Profile ID: LFUS-PAI-O-3321367