A CPU cache is a cache used by the central processing unit of a computer to reduce the average time to access memory. The cache is a smaller, faster memory which stores copies of the data from the most frequently used main memory locations. As long as most memory accesses are to cached memory locations, the average latency of memory accesses will be closer to the cache latency than to the latency of main memory.
When the processor needs to read from or write to a location in main memory, it first checks whether a copy of that data is in the cache. If so, the processor immediately reads from or writes to the cache, which is much faster than reading from or writing to main memory.
The diagram on the right shows two memories. Each location in each memory has a datum (a cache line), which in different designs ranges in size from 8 to 512 bytes. The size of the cache line is usually larger than the size of the usual access requested by a CPU instruction, which ranges from 1 to 16 bytes. Each location in each memory also has an index, which is a unique number used to refer to that location. The index for a location in main memory is called an address. Each location in the cache has a tag that contains the index of the datum in main memory that has been cached. In a CPU's data cache these entries are called cache lines or cache blocks.
Most modern desktop and server CPUs have at least three independent caches: an instruction cache to speed up executable instruction fetch, a data cache to speed up data fetch and store, and a translation lookaside buffer used to speed up virtual-to-physical address translation for both executable instructions and data.
However, of all microprocessor units sold (including embedded processors), most of them do not have any cache—mostly to reduce cost, but sometimes to improve the determinism of a real-time computing system.
When the processor needs to read or write a location in main memory, it first checks whether that memory location is in the cache. This is accomplished by comparing the address of the memory location to all tags in the cache that might contain that address. If the processor finds that the memory location is in the cache, we say that a cache hit has occurred, otherwise we speak of a cache miss. In the case of a cache hit, the processor immediately reads or writes the data in the cache line. The proportion of accesses that result in a cache hit is known as the hit rate, and is a measure of the effectiveness of the cache.
In the case of a cache miss, most caches allocate a new entry, which comprises the tag just missed and a copy of the data from memory. The reference can then be applied to the new entry just as in the case of a hit. Misses are comparatively slow because they require the data to be transferred from main memory. This transfer incurs a delay since main memory is much slower than cache memory, and also incurs the overhead for recording the new data in the cache before it is delivered to the processor.
In order to make room for the new entry on a cache miss, the cache generally has to evict one of the existing entries. The heuristic that it uses to choose the entry to evict is called the replacement policy. The fundamental problem with any replacement policy is that it must predict which existing cache entry is least likely to be used in the future. Predicting the future is difficult, especially for hardware caches that use simple rules amenable to implementation in circuitry, so there are a variety of replacement policies to choose from and no perfect way to decide among them. One popular replacement policy, LRU, replaces the least recently used entry.
When data is written to the cache, it must at some point be written to main memory as well. The timing of this write is controlled by what is known as the write policy. In a write-through cache, every write to the cache causes a write to main memory. Alternatively, in a write-back or copy-back cache, writes are not immediately mirrored to memory. Instead, the cache tracks which locations have been written over (these locations are marked dirty). The data in these locations is written back to main memory when that data is evicted from the cache. For this reason, a miss in a write-back cache will often require two memory accesses to service: one to first write the dirty location to memory and then another to read the new location from memory.
There are intermediate policies as well. The cache may be write-through, but the writes may be held in a store data queue temporarily, usually so that multiple stores can be processed together (which can reduce bus turnarounds and so improve bus utilization).
The data in main memory being cached may be changed by other entities, in which case the copy in the cache may become out-of-date or stale. Alternatively, when the CPU updates the data in the cache, copies of data in other caches will become stale. Communication protocols between the cache managers which keep the data consistent are known as cache coherence protocols.
The time taken to fetch a datum from memory (read latency) matters because a CPU will often run out of things to do while waiting for the datum. When a CPU reaches this state, it is called a stall. As CPUs become faster, stalls due to cache misses displace more potential computation; modern CPUs can execute hundreds of instructions in the time taken to fetch a single datum from memory. Various techniques have been employed to keep the CPU busy during this time. Out-of-order CPUs (Pentium Pro and later Intel designs, for example) attempt to execute independent instructions after the instruction that is waiting for the cache miss data. Another technology, used by many processors, is simultaneous multithreading (SMT), or in Intel's terminology hyper-threading (HT), which allows an alternate thread to use the CPU core while a first thread waits for data to come from main memory.
Cache row entries usually have the following structure:
|Data blocks||Tag||Index||Displacement||Valid bit|
The data blocks contain the actual data fetched from the main memory. The memory address is split (MSB to LSB) into the tag, the index and the displacement, while the valid bit denotes that this particular entry has valid data. The index length is bits and describes which row the data has been put in. The displacement length is and specifies which block of the ones we have stored we need. The tag length is and contains the most significant bits of the address, which are checked against the current row (the row has been retrieved by index) to see if it is the one we need or another, irrelevant memory location that happened to have the same index bits as the one we want.
The replacement policy decides where in the cache a copy of a particular entry of main memory will go. If the replacement policy is free to choose any entry in the cache to hold the copy, the cache is called fully associative. At the other extreme, if each entry in main memory can go in just one place in the cache, the cache is direct mapped. Many caches implement a compromise in which each entry in main memory can go to any one of N places in the cache, and are described as N-way set associative. For example, the level-1 data cache in an AMD Athlon is 2-way set associative, which means that any particular location in main memory can be cached in either of 2 locations in the level-1 data cache.
Associativity is a trade-off. If there are ten places the replacement policy can put a new cache entry, then when the cache is checked for a hit, all ten places must be searched. Checking more places takes more power, area, and potentially time. On the other hand, caches with more associativity suffer fewer misses (see conflict misses, below), so that the CPU spends less time servicing those misses. The rule of thumb is that doubling the associativity, from direct mapped to 2-way, or from 2-way to 4-way, has about the same effect on hit rate as doubling the cache size. Associativity increases beyond 4-way have much less effect on the hit rate, and are generally done for other reasons (see virtual aliasing, below).
In order of increasing (worse) hit times and decreasing (better) miss rates,
If each location in main memory can be cached in either of two locations in the cache, one logical question is: which two? The simplest and most commonly used scheme, shown in the right-hand diagram above, is to use the least significant bits of the memory location's index as the index for the cache memory, and to have two entries for each index. One good property of this scheme is that the tags stored in the cache do not have to include that part of the main memory address which is implied by the cache memory's index. Since the cache tags are fewer bits, they take less area [on the microprocessor chip] and can be read and compared faster.
One of the advantages of a direct mapped cache is that it allows simple and fast speculation. Once the address has been computed, the one cache index which might have a copy of that datum is known. That cache entry can be read, and the processor can continue to work with that data before it finishes checking that the tag actually matches the requested address.
The idea of having the processor use the cached data before the tag match completes can be applied to associative caches as well. A subset of the tag, called a hint, can be used to pick just one of the possible cache entries mapping to the requested address. This datum can then be used in parallel with checking the full tag. The hint technique works best when used in the context of address translation, as explained below.
Other schemes have been suggested, such as the skewed cache, where the index for way 0 is direct, as above, but the index for way 1 is formed with a hash function. A good hash function has the property that addresses which conflict with the direct mapping tend not to conflict when mapped with the hash function, and so it is less likely that a program will suffer from an unexpectedly large number of conflict misses due to a pathological access pattern. The downside is extra latency from computing the hash function. Additionally, when it comes time to load a new line and evict an old line, it may be difficult to determine which existing line was least recently used, because the new line conflicts with data at different indexes in each way; LRU tracking for non-skewed caches is usually done on a per-set basis. Nevertheless, skewed-associative caches have major advantages over conventional set-associative ones.
In the common case of finding a hit in the first way tested, a pseudo-associative cache is as fast as a direct-mapped cache. But it has a much lower conflict miss rate than a direct-mapped cache, closer to the miss rate of a fully associative cache.
A cache read miss from an instruction cache generally causes the most delay, because the processor, or at least the thread of execution, has to wait (stall) until the instruction is fetched from main memory.
A cache read miss from a data cache usually causes less delay, because instructions not dependent on the cache read can be issued and continue execution until the data is returned from main memory, and the dependent instructions can resume execution.
A cache write miss to a data cache generally causes the least delay, because the write can be queued and there are few limitations on the execution of subsequent instructions. The processor can continue until the queue is full.
In order to lower cache miss rate, a great deal of analysis has been done on cache behavior in an attempt to find the best combination of size, associativity, block size, and so on. Sequences of memory references performed by benchmark programs are saved as address traces. Subsequent analyses simulate many different possible cache designs on these long address traces. Making sense of how the many variables affect the cache hit rate can be quite confusing. One significant contribution to this analysis was made by Mark Hill, who separated misses into three categories (known as the Three Cs):
The graph to the right summarizes the cache performance seen on the Integer portion of the SPEC CPU2000 benchmarks, as collected by Hill and Cantin These benchmarks are intended to represent the kind of workload that an engineering workstation computer might see on any given day. The reader should keep in mind that finding benchmarks which are even usefully representative of many programs has been very difficult, and there will always be important programs with very different behavior than what is shown here.
We can see the different effects of the three Cs in this graph.
At the far right, with cache size labelled "Inf", we have the compulsory misses. If we wish to improve a machine's performance on SpecInt2000, increasing the cache size beyond 1 MiB is essentially futile. That's the insight given by the compulsory misses.
The fully-associative cache miss rate here is almost representative of the capacity miss rate. The difference is that the data presented is from simulations assuming an LRU replacement policy. Showing the capacity miss rate would require a perfect replacement policy, i.e. an oracle that looks into the future to find a cache entry which is actually not going to be hit.
Note that our approximation of the capacity miss rate falls steeply between 32 KiB and 64 KiB. This indicates that the benchmark has a working set of roughly 64 KiB. A CPU cache designer examining this benchmark will have a strong incentive to set the cache size to 64 KiB rather than 32 KiB. Note that, on this benchmark, no amount of associativity can make a 32 KiB cache perform as well as a 64 KiB 4-way, or even a direct-mapped 128 KiB cache.
Finally, note that between 64 KiB and 1 MiB there is a large difference between direct-mapped and fully-associative caches. This difference is the conflict miss rate. The insight from looking at conflict miss rates is that secondary caches benefit a great deal from high associativity.
This benefit was well known in the late 80s and early 90s, when CPU designers could not fit large caches on-chip, and could not get sufficient bandwidth to either the cache data memory or cache tag memory to implement high associativity in off-chip caches. Desperate hacks were attempted: the MIPS R8000 used expensive off-chip dedicated tag SRAMs, which had embedded tag comparators and large drivers on the match lines, in order to implement a 4 MiB 4-way associative cache. The MIPS R10000 used ordinary SRAM chips for the tags. Tag access for both ways took two cycles. To reduce latency, the R10000 would guess which way of the cache would hit on each access.
Virtual memory requires the processor to translate virtual addresses generated by the program into physical addresses in main memory. The portion of the processor that does this translation is known as the memory management unit (MMU). The fast path through the MMU can perform those translations stored in the Translation Lookaside Buffer (TLB), which is a cache of mappings from the operating system's page table. For the purposes of the present discussion, there are three important features of address translation:
A historical note: the first virtual memory systems were very slow, because they required an access to the page table (held in main memory) before every programmed access to main memory. With no caches, this effectively cut the speed of the machine in half. The first hardware cache used in a computer system was not actually a data or instruction cache, but rather a TLB.
Virtually indexed caches use a portion of the virtual address for their index, which is available earlier than the physical address. If the cache is either direct mapped or virtually tagged, there is no need to consult the MMU to determine which data to feed back to the execution datapath, and so the cache can be very fast (especially if small, as were the primary data caches in the first two generations of the Intel Pentium 4). The speed of this recurrence (the load latency) is crucial to CPU performance, and so most modern level-1 caches are virtually indexed, which at least allows the MMU's TLB lookup to proceed in parallel with fetching the data from the cache RAM.
But virtual indexing is not the best choice for all cache levels. It introduces the problem of virtual aliases—the cache may have multiple locations which can store the value of a single physical address. The cost of dealing with virtual aliases grows with cache size, and as a result most level-2 and larger caches are physically indexed.
Caches have historically used both virtual and physical addresses for the cache tags, although virtual tagging is now uncommon. If the TLB lookup can finish before the cache RAM lookup, then the physical address is available in time for tag compare, and there is no need for virtual tagging. Large caches, then, tend to be physically tagged, and only small, very low latency caches are virtually tagged. In recent general-purpose CPUs, virtual tagging has been superseded by vhints, as described below.
Whenever a new entry is added to a virtually-indexed cache, the processor searches for any virtual aliases already resident and evicts them first. This special handling happens only during a cache miss. No special work is necessary during a cache hit, which helps keep the fast path fast.
The most straightforward way to find aliases is to arrange for them all to map to the same location in the cache. This happens, for instance, if the TLB has e.g. 4 KiB pages, and the cache is direct mapped and 4 KiB or less.
Modern level-1 caches are much larger than 4 KiB, but virtual memory pages have stayed that size. If the cache is e.g. 16 KiB and virtually indexed, for any virtual address there are four cache locations that could hold the same physical location, but aliased to different virtual addresses. If the cache misses, all four locations must be probed to see if their corresponding physical addresses match the physical address of the access that generated the miss.
These probes are the same checks that a set associative cache uses to select a particular match. So if a 16 KiB virtually indexed cache is 4-way set associative and used with 4 KiB virtual memory pages, no special work is necessary to evict virtual aliases during cache misses because the checks have already happened while checking for a cache hit.
Using the AMD Athlon as an example again, it has a 64 KiB level-1 data cache, 4 KiB pages, and 2-way set associativity. When the level-1 data cache suffers a miss, 2 of the 16 (==64 KiB/4 KiB) possible virtual aliases have already been checked, and seven more cycles through the tag check hardware are necessary to complete the check for virtual aliases.
It is also possible for the operating system to ensure that no virtual aliases are simultaneously resident in the cache. The operating system makes this guarantee by enforcing page coloring, which is described below. Some early RISC processors (SPARC, RS/6000) took this approach. It has not been used recently, as the hardware cost of detecting and evicting virtual aliases has fallen and the software complexity and performance penalty of perfect page coloring has risen.
It can be useful to distinguish the two functions of tags in an associative cache: they are used to determine which way of the entry set to select, and they are used to determine if the cache hit or missed. The second function must always be correct, but it is permissible for the first function to guess, and get the wrong answer occasionally.
Some processors (e.g. early SPARCs) have caches with both virtual and physical tags. The virtual tags are used for way selection, and the physical tags are used for determining hit or miss. This kind of cache enjoys the latency advantage of a virtually tagged cache, and the simple software interface of a physically tagged cache. It bears the added cost of duplicated tags, however. Also, during miss processing, the alternate ways of the cache line indexed have to be probed for virtual aliases and any matches evicted.
The extra area (and some latency) can be mitigated by keeping virtual hints with each cache entry instead of virtual tags. These hints are a subset or hash of the virtual tag, and are used for selecting the way of the cache from which to get data and a physical tag. Like a virtually tagged cache, there may be a virtual hint match but physical tag mismatch, in which case the cache entry with the matching hint must be evicted so that cache accesses after the cache fill at this address will have just one hint match. Since virtual hints have fewer bits than virtual tags distinguishing them from one another, a virtually hinted cache suffers more conflict misses than a virtually tagged cache.
Perhaps the ultimate reduction of virtual hints can be found in the Pentium 4 (Willamette and Northwood cores). In these processors the virtual hint is effectively 2 bits, and the cache is 4-way set associative. Effectively, the hardware maintains a simple permutation from virtual address to cache index, so that no content-addressable memory (CAM) is necessary to select the right one of the four ways fetched.
To understand the problem, consider a CPU with a 1 MiB physically indexed direct-mapped level-2 cache and 4 KiB virtual memory pages. Sequential physical pages map to sequential locations in the cache until after 256 pages the pattern wraps around. We can label each physical page with a color of 0–255 to denote where in the cache it can go. Locations within physical pages with different colors cannot conflict in the cache.
A programmer attempting to make maximum use of the cache may arrange his program's access patterns so that only 1 MiB of data need be cached at any given time, thus avoiding capacity misses. But he should also ensure that the access patterns do not have conflict misses. One way to think about this problem is to divide up the virtual pages the program uses and assign them virtual colors in the same way as physical colors were assigned to physical pages before. The programmer can then arrange the access patterns of his code so that no two pages with the same virtual color are in use at the same time. There is a wide literature on such optimizations (e.g. loop nest optimization), largely coming from the High Performance Computing (HPC) community.
The snag is that while all the pages in use at any given moment may have different virtual colors, some may have the same physical colors. In fact, if the operating system assigns physical pages to virtual pages randomly and uniformly, it is extremely likely that some pages will have the same physical color, and then locations from those pages will collide in the cache (this is the birthday paradox).
The solution is to have the operating system attempt to assign different physical color pages to different virtual colors, a technique called page coloring. Although the actual mapping from virtual to physical color is irrelevant to system performance, odd mappings are difficult to keep track of and have little benefit, so most approaches to page coloring simply try to keep physical and virtual page colors the same.
If the operating system can guarantee that each physical page maps to only one virtual color, then there are no virtual aliases, and the processor can use virtually indexed caches with no need for extra virtual alias probes during miss handling. Alternatively, the O/S can flush a page from the cache whenever it changes from one virtual color to another. As mentioned above, this approach was used for some early SPARC and RS/6000 designs.
Pipelines with separate instruction and data caches, now predominant, are said to have a Harvard architecture. Originally, this phrase referred to machines with separate instruction and data memories, which proved not at all popular. Most modern CPUs have a single-memory von Neumann architecture.
The original victim cache on the HP PA7200 was a small, fully-associative cache. Later processors, such as the AMD K7 and K8, used the very large secondary cache as a victim cache, to avoid duplicate storage of the contents of the large primary cache.
The earliest widely acknowledged academic publication of trace cache was by Eric Rotenberg, Steve Bennett, and James E. Smith in their 1996 paper "Trace Cache: a Low Latency Approach to High Bandwidth Instruction Fetching."
An earlier publication is US Patent 5,381,533, "Dynamic flow instruction cache memory organized around trace segments independent of virtual address line", by Alex Peleg and Uri Weiser of Intel Corp., patent filed March 30, 1994, a continuation of an application filed in 1992, later abandoned.
A trace cache stores instructions either after they have been decoded, or as they are retired. Generally, instructions are added to trace caches in groups representing either individual basic blocks or dynamic instruction traces. A basic block consists of a group of non-branch instructions ending with a branch. A dynamic trace ("trace path") contains only instructions whose results are actually used, and eliminates instructions following taken branches (since they are not executed); a dynamic trace can be a concatenation of multiple basic blocks. This allows the instruction fetch unit of a processor to fetch several basic blocks, without having to worry about branches in the execution flow.
Trace lines are stored in the trace cache based on the program counter of the first instruction in the trace and a set of branch predictions. This allows for storing different trace paths that start on the same address, each representing different branch outcomes. In the instruction fetch stage of a pipeline, the current program counter along with a set of branch predictions is checked in the trace cache for a hit. If there is a hit, a trace line is supplied to fetch which does not have to go to a regular cache or to memory for these instructions. The trace cache continues to feed the fetch unit until the trace line ends or until there is a misprediction in the pipeline. If there is a miss, a new trace starts to be built.
Trace caches are also used in processors like the Intel Pentium 4 to store already decoded micro-operations, or translations of complex x86 instructions, so that the next time an instruction is needed, it does not have to be decoded again.
Multi-level caches generally operate by checking the smallest Level 1 (L1) cache first; if it hits, the processor proceeds at high speed. If the smaller cache misses, the next larger cache (L2) is checked, and so on, before external memory is checked.
As the latency difference between main memory and the fastest cache has become larger, some processors have begun to utilize as many as three levels of on-chip cache. For example, in 2003, Itanium 2 began shipping with a 6 MiB unified level 3 (L3) cache on-chip. The IBM Power 4 series has a 256 MiB L3 cache off chip, shared among several processors. The new AMD Phenom series of chips carries a 2MB on die L3 cache.
The advantage of exclusive caches is that they store more data. This advantage is larger when the exclusive L1 cache is comparable to the L2 cache, and diminishes if the L2 cache is many times larger than the L1 cache. When the L1 misses and the L2 hits on an access, the hitting cache line in the L2 is exchanged with a line in the L1. This exchange is quite a bit more work than just copying a line from L2 to L1, which is what an inclusive cache does.
One advantage of strictly inclusive caches is that when external devices or other processors in a multiprocessor system wish to remove a cache line from the processor, they need only have the processor check the L2 cache. In cache hierarchies which do not enforce inclusion, the L1 cache must be checked as well. As a drawback, there is a correlation between the associativities of L1 and L2 caches: if the L2 cache does not have at least as many ways as all L1 caches together, the effective associativity of the L1 caches is restricted.
Another advantage of inclusive caches is that the larger cache can use larger cache lines, which reduces the size of the secondary cache tags. (Exclusive caches require both caches to have the same size cache lines, so that cache lines can be swapped on a L1 miss, L2 hit). If the secondary cache is an order of magnitude larger than the primary, and the cache data is an order of magnitude larger than the cache tags, this tag area saved can be comparable to the incremental area needed to store the L1 cache data in the L2.
As mentioned above, larger computers sometimes have another cache between the L2 cache and main memory called an L3 cache. This cache can be implemented on a separate chip from the CPU, and, as of 2004, may range in size from 2 to 256 megabytes. The benefits of an off chip L3 cache depend on the application's access patterns. High-end x86 workstations and servers are now available with an L3 cache option implemented on the microprocessor die, increasing the speed and reducing the cost substantially. For example, Intel's Xeon MP product code-named "Tulsa" features 16 MiB of on-die L3 cache, shared between two processor cores.
Finally, at the other end of the memory hierarchy, the CPU register file itself can be considered the smallest, fastest cache in the system, with the special characteristic that it is scheduled in software—typically by a compiler, as it allocates registers to hold values retrieved from main memory. (See especially loop nest optimization.) Register files sometimes also have hierarchy: The Cray-1 (circa 1976) had 8 address "A" and 8 scalar data "S" registers that were generally usable. There was also a set of 64 address "B" and 64 scalar data "T" registers that took longer to access, but were faster than main memory. The "B" and "T" registers were provided because the Cray-1 did not have a data cache. (The Cray-1 did, however, have an instruction cache.)
The K8 has 4 specialized caches: an instruction cache, an instruction TLB, a data TLB, and a data cache. Each of these caches is specialized:
The K8 also has multiple-level caches. There are second-level instruction and data TLBs, which store only PTEs mapping 4 KiB. Both instruction and data caches, and the various TLBs, can fill from the large unified L2 cache. This cache is exclusive to both the L1 instruction and data caches, which means that any 8-byte line can only be in one of the L1 instruction cache, the L1 data cache, or the L2 cache. It is, however, possible for a line in the data cache to have a PTE which is also in one of the TLBs—the operating system is responsible for keeping the TLBs coherent by flushing portions of them when the page tables in memory are updated.
The K8 also caches information that is never stored in memory—prediction information. These caches are not shown in the above diagram. As is usual for this class of CPU, the K8 has fairly complex branch prediction, with tables that help predict whether branches are taken and other tables which predict the targets of branches and jumps. Some of this information is associated with instructions, in both the level 1 instruction cache and the unified secondary cache.
The K8 uses an interesting trick to store prediction information with instructions in the secondary cache. Lines in the secondary cache are protected from accidental data corruption (e.g. by an alpha particle strike) by either ECC or parity, depending on whether those lines were evicted from the data or instruction primary caches. Since the parity code takes fewer bits than the ECC code, lines from the instruction cache have a few spare bits. These bits are used to cache branch prediction information associated with those instructions. The net result is that the branch predictor has a larger effective history table, and so has better accuracy.
These predictors are caches in that they store information that is costly to compute. Some of the terminology used when discussing predictors is the same as that for caches (one speaks of a hit in a branch predictor), but predictors are not generally thought of as part of the cache hierarchy.
The K8 keeps the instruction and data caches coherent in hardware, which means that a store into an instruction closely following the store instruction will change that following instruction. Other processors, like those in the Alpha and MIPS family, have relied on software to keep the instruction cache coherent. Stores are not guaranteed to show up in the instruction stream until a program calls an operating system facility to ensure coherency. The idea is to save hardware complexity on the assumption that self-modifying code is rare.
The simplest cache is a virtually indexed direct-mapped cache. The virtual address is calculated with an adder, the relevant portion of the address extracted and used to index an SRAM, which returns the loaded data. The data is byte aligned in a byte shifter, and from there is bypassed to the next operation. There is no need for any tag checking in the inner loop — in fact, the tags need not even be read. Later in the pipeline, but before the load instruction is retired, the tag for the loaded data must be read, and checked against the virtual address to make sure there was a cache hit. On a miss, the cache is updated with the requested cache line and the pipeline is restarted.
An associative cache is more complicated, because some form of tag must be read to determine which entry of the cache to select. An N-way set-associative level-1 cache usually reads all N possible tags and N data in parallel, and then chooses the data associated with the matching tag. Level-2 caches sometimes save power by reading the tags first, so that only one data element is read from the data SRAM.
The diagram to the right is intended to clarify the manner in which the various fields of the address are used. Address bit 31 is most significant, bit 0 is least significant. The diagram shows the SRAMs, indexing, and multiplexing for a 4 KiB, 2-way set-associative, virtually indexed and virtually tagged cache with 64 B lines, a 32b read width and 32b virtual address.
Because the cache is 4 KiB and has 64 B lines, there are just 64 lines in the cache, and we read two at a time from a Tag SRAM which has 32 rows, each with a pair of 21 bit tags. Although any function of virtual address bits 31 through 6 could be used to index the tag and data SRAMs, it is simplest to use the least significant bits.
Similarly, because the cache is 4 KiB and has a 4 B read path, and reads two ways for each access, the Data SRAM is 512 rows by 8 bytes wide.
A more modern cache might be 16 KiB, 4-way set-associative, virtually indexed, virtually hinted, and physically tagged, with 32 B lines, 32b read width and 36b physical addresses. The read path recurrence for such a cache looks very similar to the path above. Instead of tags, vhints are read, and matched against a subset of the virtual address. Later on in the pipeline, the virtual address is translated into a physical address by the TLB, and the physical tag is read (just one, as the vhint supplies which way of the cache to read). Finally the physical address is compared to the physical tag to determine if a hit has occurred.
Some SPARC designs have improved the speed of their L1 caches by a few gate delays by collapsing the virtual address adder into the SRAM decoders. See Sum addressed decoder.
The arrival of PCs coincided with a temporary decline in interest in caching. In the early days of PC technology, memory access was only slightly slower than register access. But since the 1980s the performance gap between processor and memory has been growing. Processors have advanced much faster than memory, especially in terms of their operating frequency, so memory became a performance bottleneck. While it was technically possible to have all the main memory as fast as the processor, a more economically viable path has been taken: use plenty of low-speed memory, but also introduce a small high-speed cache memory to alleviate the performance gap. This provided an order of magnitude more capacity—for the same price—with only a slightly-reduced combined performance.
Some versions of the Intel 386 type processor could support 16 to 64 KiB of external cache.
With the 486 processor an 8 KiB cache was integrated directly into the CPU die. This cache was termed Level 1 or L1 cache to differentiate it from the slower on-motherboard, or Level 2 (L2) cache. These on-motherboard caches were much larger, with the most common size being 256 KiB and frequently utilizing a SIMM form factor. The popularity of on-motherboard cache continued on through the Pentium MMX era but was made obsolete by the introduction of SDRAM and the growing disparity between bus speed and CPU clock speed, which caused on-motherboard cache to be only slightly faster than main memory.
The next evolution in cache implementation in the x86 architecture begun with the Pentium Pro, which brought the secondary cache onto the same package as the microprocessor, clocked at same frequency as the microprocessor.