The processors of the Intel Coffee Lake family remain one of the most “hot” topics of this season. This is clearly visible from the current range of retail stores. Although the announcement of Intel’s new products took place more than two months ago, on the shelves they are presented somewhat limited, and the prices for existing models due to the deficit are far from the recommended values. But despite the significant shortages, the older Coffee Lake easily fall into the number of bestsellers – just look at the sales level. For example, in the statistics of the largest online store Amazon.com Core i7-8700K confidently takes the first place among the processors among processors more than $ 300.
With the appearance of the Coffee Lake design in the older series of Intel processors, the number of computing cores increased one and a half times, which inevitably and very significantly affected the performance. However, do not forget about the existence of the younger series of Core i3, with which even more serious changes occurred. While until recently only processors with two processing cores were included in it, Core i3 now became full-fledged quad-core, but without the support of Hyper-Threading technology. This means that Intel has increased the power of its inexpensive series even more than the older chips, and such a change clearly deserves a separate study.
At first glance, the new Core i3 looks just fine. By basic characteristics, they are very similar to the Core i5 of the Kaby Lake era, but they have a price below $ 170 (in theory), which is traditionally typical for the Core i3 series. However, in practice the situation is somewhat different, and the older Core i3-8350K is currently selling 10-15 percent more expensive than the older Core i3-7350K of the previous generation. The result is that, contrary to the original design, price parallels have to be made between the Intel Core i3-8350K and Core i5-7400 quad-core processors, but even with this positioning, the new Core i3 looks more interesting. They not only offer higher frequencies, but they also have a Core i3-8350K version with an unlocked multiplier in their model range, which is not the cheapest Intel CPU with overclocking, but the cheapest overclocker quad core even with overpriced prices.
That is why we decided to devote our next review to the Core i3-8350K processor, which can become both an excellent basis for a relatively inexpensive system and an appropriate solution for budget-minded enthusiasts who are not ready to pay over 200 USD for open overclocker functions.
However, to draw any serious conclusions about the Core i3-8350K without a detailed investigation would not be very correct. For example, the previous processor of this kind, Core i3-7350K, has not become a popular model, despite the moderate price and overclocking possibilities. The new Core i3-8350K offers a much more advantageous combination of performance and price, but one can also doubt its unequivocal superiority in its price category. The fact is that the processors with six processing cores cost only 10-20 dollars more, and the Core i5-8400 or Ryzen 5 1600 can be more profitable alternatives to the Core i3-8350K. To understand how it really is, we decided in this review.
Core i3-8350K in detail
The processors of the generation of Coffee Lake in the microarchitecture do not differ from the representatives of the Kaby Lake family and represent an update of the previous design, made in the light of the introduction of the improved technological process of 14 ++ nm. This means that, given the equality of the number of cores, operating frequencies and other characteristics, Coffee Lake and Kaby Lake will demonstrate identical performance. Thus, if you disregard the issues of positioning and pricing, the new Core i3-8350K is logical to compare with the processor Core i5-7600K generation Kaby Lake.
And really: both Core i3-8350K and Core i5-7600K are very similar quad-core without Hyper-Threading technology support, but with an unlocked multiplier that allows them to be overclocked on motherboards based on Z-series system logic sets. True, the Core i3-8350K is formally designed for slightly lower clock speeds, and besides, like all other representatives of the Core i3 series, it does not support the automatic overclocking technology of Turbo Boost. But the free multiplier easily compensates for all the frequency limitations in the specifications Core i3-8350K, which makes it equivalent in features and more affordable alternative to the older Core i5 generation Kaby Lake.
In support of what has been said, here is a table that compares the formal specifications of the Core i3-8350K and Core i5-7600K. In addition, in the same table we also placed the third processor – Core i5-8400. Although formally he refers to the heavier artillery, since he has six cores, Core i5-8400 will most likely be the main competitor for Core i3-8350K. Intel’s price list assumes a very ephemeral boundary between the processor families Core i3 and Core i5. The junior six-core is only $ 14 more expensive than the older quad core, and this is hardly a principal amount. Therefore, when choosing between Core i3-8350K and Core i5-8400, the main factor will be what is more important in each case: an additional two cores or overclocking capability.
|Core i5-8400||Core i5-7600K||Core i3-8350K|
|Code Name||Coffee Lake||Kaby Lake||Coffee Lake|
|Production technology, nm||14++||14+||14++|
|Kernels / streams||6/6||4/4||4/4|
|Base frequency, GHz||2,8||3,8||4,0|
|Frequency Turbo Boost 2.0, GHz||4,0||4,2||–|
|Integrated graphics||GT2: 24 EU||GT2: 24 EU||GT2: 23 EU|
|Max. frequency of the graphics core, GHz||1,05||1,15||1,15|
|PCI Express Lines||16||16||16|
|Socket||LGA1151 v2||LGA1151 v1||LGA1151 v2|
The lack of support for Turbo Boost technology in Core i3-8350K means that it always works at the same clock speed – 4.0 GHz, regardless of the load. The Core i5-7600K, whose base frequency is 200 MHz lower, actually operates at a frequency of 4.0 to 4.2 GHz, which varies depending on the load under the Turbo Boost technology. However, this does not mean that Core i3-8350K will lag behind the older Core i5 of the previous generation in real-world tasks.
The Core i3-8350K has an important advantage of a different kind – a more capacious third-level cache, the full size of which is formed in the same way as in the Core i7 processors: based on 2 MB per core. At processors Core i5 both past, and present generation capacity L3-cache makes only on 1,5 Mbyte on a core therefore as to the size of a cache-memory Core i3-8350K surpasses Core i5-7600K on third. And moreover, the Core i3-8350K cache memory is almost as capacious as the six-core Core i5 generation of Coffee Lake.
However, the fact that Intel is generous for Core i3-8350K on the cache memory does not at all mean that this processor has no limitations in other capabilities. For example, only DDR4-2400 is declared for it, while the older series processors are officially compatible with DDR4-2666. However, high-speed memory modules without any problems are supported by the new quad core through overclocking – so there is no trick here. But what can not be fixed with simple settings is that one of the 24 executive devices is disabled in the Core i3 graphics core, because of which the integrated graphics Core i3-8350K, although called UHD Graphics 630, works a little slower than the more dear brothers.
Comparing the Core i3-8350K with the Core i5-7600K seems quite logical, but a couple of words worth mentioning about how the characteristics of the new quad core i3 correlate with what the Core i3-7350K, which is related to the Kaby Lake generation, offered. Yes, the performance of the predecessor Core i3-8350K, obviously, is noticeably smaller, because it has only two computational cores, albeit with support for Hyper-Threading technology. The performance gap, which is provided by additional high-grade Core i3-8350K cores, can not be compensated at 200 MHz by higher Core i3-7350K frequencies.
But according to one characteristic, the Core i3-7350K still wins from its quad-core follower. Dual-core overclocker Kaby Lake was a 60-watt thermal package, that is, it was noticeably more economical than the Core i3-8350K, the rated heat dissipation of which reaches 91 watts. Energy efficiency does not belong to the strengths of Coffee Lake’s design, and that’s another confirmation of that. According to the consumption and heat release of Core i3-8350K, which is based on a 14-nm semiconductor chip manufactured according to the technical process, it is supplied by the manufacturer in a row with the four-core processors of Kaby Lake, manufactured under the previous version of the 14-nm process technology.
Core i3-8350K vs Core i5-7600K: L3 cache comparison
If you carefully read the last section, you probably understood that the only architectural feature of the Core i3-8350K, which deserves detailed consideration, is its third-level cache memory. Not that the L3-cache in the four-core generations of the Coffee Lake generation was somehow different from the cache memory of similar processors with the Kaby Lake design, but there is a significant difference in capacity: Core i3-8350K got an 8-MB L3-cache, at that time as the quad-core Core i5 of the previous generation had a 6-megabyte cache memory of the third level. The difference in size also causes a different associative policy: for the Core i3-8350K, the L3 cache has a maximum of 16-channel associativity for the Coffee Lake, while the Kaby Lake processors have a 12-channel associative memory cache for Core i5 processors. The latency of the cache memory is the same in both cases (42 bars), so in general it can be said that the Core i3-8350K has not only a more capacious, but also a more efficient caching system.
In most other characteristics, the cache of the third level Core i3-8350K is identical to the Core i5-7600K cache. There are no differences even in its frequency of work:
|Core i5-7600K||Core i3-8350K|
|L3 Cache Characteristics|
|Size L3-cache, MB||6||8|
|Line size, byte||64||64|
|Write policy||Postponed (write-back)||Postponed (write-back)|
|Tire width, byte||32||32|
All the above is easy to illustrate with the results of practical measurements, which do not show any fundamental difference in the cache memory speed for both processors. Thus, the graphs below demonstrate the latency of the cache memory subsystem, measured by the SiSoftware Sandra utility, with the help of the utility, when working with data blocks of various sizes. For clarity, the frequency of Core i3-8350K and Core i5-7600K during this test was fixed at the same value of 4.0 GHz.
The latency of L1, L2 and L3 caches in Core i3-8350K and Core i5-7600K is the same. However, the lines corresponding to these processors coincide on the charts incompletely, which is caused by the more capacious L3-cache in the new four-core Core i3.
In addition, some differences can be seen in the memory latency. I must say that, as shown in more detailed testing in SiSoftware Sandra, the memory controllers Core i3-8350K and Core i5-7600K are by no means identical in performance. Coffee Lake works with memory a little slower – this is indicated by the results of measurements that we carried out in both systems with the same DDR4-3000 modules with 16-16-16-36-1T delay scheme. Perhaps this is due to the increased omnivorousness of the new controller, which has become compatible with a wider range of high-speed DDR4 SDRAM modules.
However, the differences in the speed of memory subsystems Core i3-8350K and Core i5-7600K can not be called any fundamental. Nevertheless, it is already impossible to write off the slightly worse characteristics of the Coffee Lake memory controller on the immaturity of the LGA1151 platform of the second version. Since the release of this platform has passed enough time, and optimization of performance parameters in the boards based on the Z370 logic set has now been completed.