iXBT Labs - Computer Hardware in Detail






Intel Core i5, Core i7 LGA1156 Processors

<< Previous page

     Next page >>

Testbed configurations

Processor Core i5 750 Core i7 860 Core i7 870
Core name Lynnfield
Process technology, nm 45
Core clock (std/max), GHz 2.66/3.2 2.8/3.47 2.93/3.6
Cores/HT threads 4/4 4/8
L1 cache, I/D, KB 32/32
L2 Cache, KB 4 x 256
L3 cache, KB 8192
Memory 2 x DDR3-1333
Multiplier 20 21 22
QPI 4.8 GT/s
Socket LGA1156
TDP, W 95

As we can see, clock rates (both the starting and the boost ones) in our three processors differ by one step sharp like in the Xeon family, but not in the Core i7 900, where the product range is more "lax" (by the way, the naming system depending on the clock rate has also been copied from Xeons). Besides, the lower model lacks HT support and has lower UnCore frequency (we've already mentioned it above). There are no other differences.

Processor Core i7 920 Core i7 950 Core i7 Extreme 975 Xeon X5560 Core 2 Quad Q9650 Phenom II X4 965
Core name Bloomfield Yorkfield Deneb
Process technology, nm 45
Core clock (std/max), GHz 2.66/2.93 3.06/3.33 3.33/3.6 2.8/3.2 3.0 3.4
Number of cores/threads 4/8 4/4
L1 cache, I/D, KB 32/32 64/64
L2 Cache, KB 4 x 256 2 x 6144 4 x 512
L3 cache, KB 8192 - 6144
Memory 3 x DDR3-1066 3 x DDR3-1333 - 2 x DDR3-1333
Multiplier 20 23 25 21 9 17
QPI 4.8 GT/s 6.4 GT/s -
Socket LGA1366 LGA775 AM2+/AM3
TDP, W 130 95 140

As it's a beginning of the new platform, there will be more competitors than models under review. However, we can easily do it without any temporary problems, as we use the same test procedure for all tests -- we've already tested all these processors. So, with what processors shall we compare our new CPUs? For one, it's the entire Core i7 family for LGA1366: 920, 950, and 975EE. For two, two guests from the other teams -- the "outdated", but still available Core 2 Quad Q9650 and the new Phenom II X4 965. You can compare the new processors with slower representative of these series on your own -- they are all included into the summary table. For three, we decided to make this review more interesting by publishing results of the Xeon X5560: it has the same starting clock rate of its cores, a tad higher UnCore frequency, the same TDP, but a tad less aggressive boost than the Core i7 860. These processors are not direct competitors, of course, but it will be interesting to compare them.

  Motherboard Memory
LGA775 ASUS P5Q3 (P45) Kingston KVR1333D3N9K3/6G (2 x 1333; 9-9-9-24-2T)
LGA1366 Intel DX58SO (X58) Kingston KVR1333D3N9K3/6G (3 x 1066; 8-8-8-19)
LGA1156 Gigabyte P55-UD6 (P55) Kingston KVR1333D3N9K3/6G (2 x 1333; 9-9-9-24)
AM3 ASUS M4A78T-E (790GX) Corsair CM3X2G1600C9DHX (2 x 1333; 7-7-7-20-1T, Unganged Mode)

All tests were performed with DDR3 memory, Intel processors were tested with the same modules. Only the number of modules were different: three modules for LGA1366 and two for LGA775 and LGA1156. We could have used configuration 2x2 for the 1366 as well. We already know that it changes test results. However, this situation would have been rather artificial. Anyway, we included results of the Core i7 920 with two memory modules into the summary table. So if you are interested in CPU performance in certain applications, you can compare processors in such conditions as well. And in the nearest future we'll try to dot all the i's for the memory configuration in the LGA1156 platform as well.


The benchmarking procedure (the list of software and test conditions) is described here. To make the diagrams easier to read, results are represented in percents (100% stands for the result of Intel Core 2 Quad Q9300 in each test). Detailed results in absolute values are published in this Microsoft Excel spreadsheet.

3D visualization

Tests of this group cannot boast of good parallelism, doing fine with two cores. So the clock rate of each core, the speed of memory operations, and other architectural peculiarities have the maximum value here, while Hyper-Threading often deteriorates performance. We can see that all three processors under review cannot boast of records, but they look very well. At least it's true for Core i5 750 and Core i7 860 -- owing to the aggressive turbo-boost and fast memory exchange they outperform their direct competitors. An interesting observation: the lack of HT served the 750 processor the good turn, it almost catches up with the 860 (with a higher starting clock rate) and defeated the equally-clocked 920. Core i7 870 is a tad slower than Core i7 950. However, the former still has such an important advantage as the total cost of the platform. Besides, the 950 has a higher nominal frequency, and the 870 is outperformed only insignificantly -- owing to the aggressive TB. What concerns processors of the previous generation as well as the AMD product, they are outscored even by the cheapest new processor.

3D rendering

People still argue whether Hyper-Threading is useful or not. As we can see, in such serious tasks as rendering you can hardly overestimate capacity of a processor to execute more threads -- performance gains are very high. So different positioning of Core i5 and i7 is perfectly justified: Model 750 and 860 differ only by 5% in frequency. What concerns performance, they differ by 20%. However, the 750 does its job well, demonstrating performance similar to the best (out of non-extreme) representative of Core 2 Quad, which is more expensive in all senses of the word: it has a higher price tag and its manufacturing costs are higher. Thus, Q9x50 processors can be painlessly discontinued now. And not to offend those users who already invested into LGA775, it will be enough to manufacture equally-clocked Q9x00 models with smaller caches and thus much cheaper. Such steps are actually already being taken -- Intel already mentioned the Q9505 with the clock rate of the Q9550, but equipped with 3M cache per core. Now it's turn for an inexpensive (its price must be at least similar to that of the Core i5 750) replacement for the Q9650. However, it's not that necessary -- Intel is not interested in prolonging the life of LGA775.

In the higher segment everything is as expected: the 860 is faster than the Core i7 920 (which renders the latter practically useless, provided their prices are equal) and Phenom II X4 965 (with their prices similar and lower power consumption of the new processor from Intel, it's a nasty situation for AMD). And the 870 is outperformed by the 950, requiring in return lower investments into the platform.

Write a comment below. No registration needed!

<< Previous page

Next page >>

blog comments powered by Disqus

  Most Popular Reviews More    RSS  

AMD Phenom II X4 955, Phenom II X4 960T, Phenom II X6 1075T, and Intel Pentium G2120, Core i3-3220, Core i5-3330 Processors

Comparing old, cheap solutions from AMD with new, budget offerings from Intel.
February 1, 2013 · Processor Roundups

Inno3D GeForce GTX 670 iChill, Inno3D GeForce GTX 660 Ti Graphics Cards

A couple of mid-range adapters with original cooling systems.
January 30, 2013 · Video cards: NVIDIA GPUs

Creative Sound Blaster X-Fi Surround 5.1

An external X-Fi solution in tests.
September 9, 2008 · Sound Cards

AMD FX-8350 Processor

The first worthwhile Piledriver CPU.
September 11, 2012 · Processors: AMD

Consumed Power, Energy Consumption: Ivy Bridge vs. Sandy Bridge

Trying out the new method.
September 18, 2012 · Processors: Intel
  Latest Reviews More    RSS  

i3DSpeed, September 2013

Retested all graphics cards with the new drivers.
Oct 18, 2013 · 3Digests

i3DSpeed, August 2013

Added new benchmarks: BioShock Infinite and Metro: Last Light.
Sep 06, 2013 · 3Digests

i3DSpeed, July 2013

Added the test results of NVIDIA GeForce GTX 760 and AMD Radeon HD 7730.
Aug 05, 2013 · 3Digests

Gainward GeForce GTX 650 Ti BOOST 2GB Golden Sample Graphics Card

An excellent hybrid of GeForce GTX 650 Ti and GeForce GTX 660.
Jun 24, 2013 · Video cards: NVIDIA GPUs

i3DSpeed, May 2013

Added the test results of NVIDIA GeForce GTX 770/780.
Jun 03, 2013 · 3Digests
  Latest News More    RSS  

Platform  ·  Video  ·  Multimedia  ·  Mobile  ·  Other  ||  About us & Privacy policy  ·  Twitter  ·  Facebook

Copyright © Byrds Research & Publishing, Ltd., 1997–2011. All rights reserved.