With the advancement of computer technology, the number of specifications for each component in a computer has become overwhelming for those not deeply involved in the computer industry. We often get questions such as "what does the frequency mean?" or "are more CPU cores always better?" In this article, we will explain the major specifications for CPUs and what they mean for the end user.
If you are interested in learning about what the specifications mean on other components, also check out out Specs Explained: Video Card article.
The product line is a group of products aimed towards a similar purpose. For example: Xeon and Opteron processors are designed for servers while Core i5 Mobile processors are designed for mobile devices such as laptops. At the time of this article, the major product lines available through Puget Systems are:
Some product lines require their own motherboard socket (such as the AMD Llano requiring socket FM1) but some can use the same socket (such as the Intel Core i3 and i5 both primarily using socket 1155).
The code name of a CPU is what further divides a product line. When a new advancement is made to the basic architecture of a CPU line, it is given a new code name. If the changes are large enough (such as a new socket design), a new product line is generally launched as well. Before a product is released, it is often referred to according to it's code name such as Sandy Bridge or Llano.
The CPU socket is the physical connection between the CPU and the motherboard. Some naming schemes are very straight forward in that the socket name is simply the number of pins between the CPU and the motherboard. For example, the Intel socket 1155 has 1155 pins, while socket 2011 has 2011 pins. Others are more confusing in their naming scheme such as the AMD socket AM3+. In those cases, the socket name doesn't give you any information as to the actual specification of the socket.
The Process specification of a CPU indicated how tightly packed the individual microscopic components (such as transistors) within the CPU can be placed when it is being manufactured. This is reported in nanometers and refers to half the distance between memory cells within the CPU itself. Many CPUs on the market today use a 32nm manufacturing process, which is over 30x smaller than the diameter of a human hair.
Smaller manufacturing processes give two major benefits: First, a smaller process means that more components can fit within a certain space. This is why CPUs today are much smaller than they were several years ago, yet are much more powerful. Second, a CPU will run cooler and require less wattage with a smaller manufacturing process. This is due to the fact that as an object becomes smaller, it becomes easier to radiate heat since the ratio of surface area to volume becomes larger. This is why heatsinks use a large number of fins in order to maximize the amount of surface are for cooling.
Number of Cores
Originally, CPUs were manufactured with a single core that did all of the work. As computers became more advanced, it was found that a single core could no longer handle the workload which resulted in the CPU becoming a bottleneck. The solution was simple: add more cores! Each core is basically a whole new processor, so by adding a second core the power of the CPU was essentially doubled. While multiple cores almost immediately had an impact on multitasking, multiple cores were inefficient at first at powering high-load applications. This was due to the fact that the algorithms used by software at the time were not very effective at utilizing multiple cores. While still not perfect, more and more software is taking advantage of the multiple cores now available within CPUs resulting in better and better performance.
The clock speed (commonly referred to as the frequency) of a CPU is how many instructions per second it can process and is typically reported in MHz or GHz. For example, a 3GHz (or 3000MHz) processor can complete 3,000,000,000 instructions per second. For example, lets create a theoretical job that requires 1,000,000,000 instructions. A 3GHz process can complete this in roughly a third of a second, while a 1GHz CPU will take a full second to complete.
Due to variances in CPU architecture, two processors with the same clock speed will not necessarily perform the same job in the same amount of time. This is why AMD and Intel processors do not always perform the same even if the number of cores and clock speed are identical. Depending on the architecture, either one CPU or the other will have an advantage due to how efficiently the CPU can process all of the instructions. This is also evident when looking at newer versus older generations of CPUs. The Intel Core i7 Extreme QUAD CORE 965 3.2GHz is much faster than the Intel Core 2 Extreme QUAD CORE QX9775 3.2GHz even though it has the same clock speed and number of cores due to advancements found in the newer CPU's architecture.
With all other things being equal - number of cores, cache size, architecture, etc - a higher clock speed will always be able to complete a set of instructions faster than a lower clock speed.
Hyperthreading is how Intel increases multitasking efficiency on many of their processors. Essentially what it does is makes your computer think that there are twice as many cores on the CPU than there actually are. So if your CPU supports hyperthreading and has 4 cores, it appears to the computer like there are actually 8 cores.
Processors are not very good at doing multiple things at once. In fact, unless the software is specially coded, a CPU will always finish the job it's working on before it moves on to a different job even if it has to pause in the middle of that job for some reason. Hyperthreading allows a core to run a second job if there is ever any downtime in the first job. The total number of instructions the core can run per second is not increased however, so the benefit of hyperthreading is only apparent when the core is not already running at 100%
Depending on what you are doing, this can either increase performance by a decent amount, or can do almost nothing. If you are just doing normal day-to-day tasks like browsing the web or checking your email, hyperthreading will be of little benefit since the CPU is already plenty fast to run all of those applications at once. Once you start doing large numbers of small things however, hyperthreading becomes useful.
Turbo Boost is available on many Intel CPUs and is essentially a temporary overclock that increases the CPU's frequency when additional processing power is needed. This can only happen as long as it the CPU is below a certain power, current and temperature threshold so it is not a full-time performance boost. The amount of frequency increase will vary depending on the number of cores that are in use. When fewer cores are in use, the clock speed of the remaining cores can be increased more than when all of the cores are in use. Since the amount of speed increase varies, we list the maximum turbo boost frequency that a processor supports when a single core is active.
Turbo Core is AMDs version of Intel's Turbo Boost; overclocking the CPU according to the workload. While the end result is similar, there are a few significant differences. Whereas Intel's Turbo Boost can increase either a single core by a large amount, or all the cores by a small amount, Turbo Core can only increase a specific number of cores. This number is exactly half the total number of cores, so a six core CPU can have three cores turbo while a four core CPU can only have two.
When the processor detects that only half the cores are being utilized, it will place the unused cores in a lower power mode (with a reduced clock speed) and increase the clock speed to the used cores. This gives a boost to the active cores allowing them to work much faster than they would at stock speeds.
The bus type of a CPU is the way in which the CPU cores communicate with the rest of the system. For the average user, the bus type will not heavily influence the speed of the processor, but newer bus types are generally more efficient than older types. At the moment, QPI (Quick Path Interconnect) is the most common bus for Intel CPUs and Hypertransport is the most common for AMD CPUs.
Thermal output (also called TDP or thermal design power) is the maximum amount of power the cooling system in the computer needs to dissipate. The higher the thermal output, the hotter the processor will run. The heat can be dissipated using a variety of different heatsinks include traditional air coolers like the Gelid Tranquillo Rev2 or liquid coolers such as the Coolit Eco II.
One often overlooked aspect of the thermal output is how it relates to the noise levels in a system. Hotter processors require more cooling which is accomplished with either a larger heatsink or by using a higher flow fan. If you use the same heatsink that you would normally use on a hot CPU and put it on a cooler CPU (one with a lower thermal output) you can use a much slower fan which in turn means much less noise. We use this methodology in our Serenity line of systems by using the fairly large Gelid Tranquillo Rev2 on a relatively low 65 or 95 watt processor.
Cache serves essentially the same purpose as the system RAM as it is a temporary storage location for data. Since L# cache is on the CPU itself however, it is much faster for the CPU to access than the main system RAM. The amount of cache available on a CPU can impact performance very heavily especially in environments with heavy multitasking.
The cache on a CPU is divided into different levels indicating the hierarchy of access. L1 is the first place the CPU looks for data and is the smallest, but also the fastest cache level. The amount of L1 cache is generally given per core and is in the range of 32KB to 64KB per core. L2 cache is the second place that the CPU looks and while larger than L1 cache is also slightly slower. L2 cache can range anywhere from 256KB to 1MB (1024KB) per core.
The reason that you do not simply make the size of the L1 cache larger instead of adding a whole new level of cache is that the larger the cache, the longer it takes for the CPU to find the data it needs. This is also that reason that it cannot be said that the more L2 cache the better. In a focused environment with only a few applications running, to a certain extent, the more cache the better. Once multitasking comes into play however, the larger cache sizes will result in the CPU having to take longer to search through all of the additional cache. For this reason, it is very difficult to say whether more L2 cache is better or not as it depends heavily on the computer's intended usage.
In general however, more L2 cache is better for the average user. In specialized applications where large amounts of small data is continuously accessed (where the total data is smaller than the total L2 cache available), less L2 cache may actually have a performance advantage over more L2 cache.
L3 cache is the third level of onboard cache and as such is the third place the CPU looks for data after first looking in the L1 and L2 cache. L3 cache is much larger than L2 or L1 cache (up to 20MB on some CPUs) but is also slower. Compared to the system RAM however, it is still much faster for the CPU to access.
L3 cache is also different in that it is almost exclusively shared across all of the cores in the CPU. So if there is data in the L3 cache, it is available for all of the cores to use unlike the core-specific L1 and L2 cache. In general, L3 cache is less concerned about speed as L1 or L2 cache so in almost all instances more L3 cache is better.
Smart Cache is essentially L3 cache, but is optimized by Intel to be more efficient at sharing cache across the multiple cores in the CPU. For all practical purposes, smart cache can be thought of as being identical to L3 cache.
The graphics processor refers to the name of the built-in graphics available on the CPU. The graphics found on a CPU is generally enough for basic tasks like web browsing, watching movies, word processing and low-end gaming but is not powerful enough to run newer 3d games at anything but the lowest settings. Not all CPUs have graphics built in, so this specification is not available on all CPUs.
Graphics Core Speed
The graphics core speed is much like the main CPU clock speed in that it tells you how many instructions per second it can process. Just like the CPU clock speed, specific instruction will always require the same number of cycles so the more cycles per second the GPU can process, the faster the job can be completed. The graphics core speed is generally much lower than the main CPU clock speed due to how differently graphical operations are handled compared to normal CPU operations.
Graphics DirectX Version
The DirectX version refers to the latest DirectX version the onboard graphics is compatible with. DirectX is a collection of APIs that is used for all graphical tasks but most widely used in the development of video games. More information on DirectX can be found on the Wikipedia page here.
Graphics Cores on Die
Cores on die are small cores built into the onboard graphical unit that do a large portion of the actual rendering of graphics. Generally, the more cores on die, the more powerful the onboard graphics although the core speed also heavily influences the power of the onboard graphics.