Intel Core i7 3960X (Sandy Bridge E) Review: Keeping the High End Alive
by Anand Lal Shimpi on November 14, 2011 3:01 AM EST- Posted in
- CPUs
- Intel
- Core i7
- Sandy Bridge
- Sandy Bridge E
The Test
To keep the review length manageable we're presenting a subset of our results here. For all benchmark results and even more comparisons be sure to use our performance comparison tool: Bench.
| Motherboard: |
ASUS P8Z68-V Pro (Intel Z68) ASUS Crosshair V Formula (AMD 990FX) Intel DX79SI (Intel X79) |
| Hard Disk: |
Intel X25-M SSD (80GB) Crucial RealSSD C300 |
| Memory: | 4 x 4GB G.Skill Ripjaws X DDR3-1600 9-9-9-20 |
| Video Card: | ATI Radeon HD 5870 (Windows 7) |
| Video Drivers: | AMD Catalyst 11.10 Beta (Windows 7) |
| Desktop Resolution: | 1920 x 1200 |
| OS: | Windows 7 x64 |
Cache and Memory Bandwidth Performance
The biggest changes from the original Sandy Bridge are the increased L3 cache size and the quad-channel memory interface. We'll first look at the impact a 15MB L3 has on latency:
| Cache/Memory Latency Comparison | ||||||
| L1 | L2 | L3 | Main Memory | |||
| AMD FX-8150 (3.6GHz) | 4 | 21 | 65 | 195 | ||
| AMD Phenom II X4 975 BE (3.6GHz) | 3 | 15 | 59 | 182 | ||
| AMD Phenom II X6 1100T (3.3GHz) | 3 | 14 | 55 | 157 | ||
| Intel Core i5 2500K (3.3GHz) | 4 | 11 | 25 | 148 | ||
| Intel Core i7 3960X (3.3GHz) | 4 | 11 | 30 | 167 | ||
Cachemem shows us a 5 cycle increase in latency. Hits in L3 can take 20% longer to get to the core that requested the data, if this is correct. For small, lightly threaded applications, you may see a slight regression in performance compared to Sandy Bridge. More likely than not however, the ~2 - 2.5x increase in L3 cache size will more than make up for the added latency. Also note that despite the large cache and thanks to its ring bus, Sandy Bridge E's L3 is still lower latency than Gulftown's.
| Memory Bandwidth Comparison - Sandra 2012.01.18.10 | |||||
| Intel Core i7 3960X (Quad Channel, DDR3-1600) | Intel Core i7 2600K (Dual Channel, DDR3-1600) | Intel Core i7 990X (Triple Channel, DDR3-1333) | |||
| Aggregate Memory Bandwidth | 37.0 GB/s | 21.2 GB/s | 19.9 GB/s | ||
Memory bandwidth is also up significantly. Populating all four channels with DDR3-1600 memory, Sandy Bridge E delivered 37GB/s of bandwidth in Sandra's memory bandwidth test. Given the 51GB/s theoretical max of this configuration and a fairly standard 20% overhead, 37GB/s is just about what we want to see here.
Facebook
Twitter
RSS
163 Comments
View All Comments
mino - Monday, November 14, 2011 - link
"Quick Sync leverages the GPU's shader array"This is simply not true. And you know it. Shame.
Steelski - Tuesday, November 15, 2011 - link
irrelevant CS4 test because someone buying this kind of hardware would appreciate the CS5 advantage other websites show.jewie27 - Tuesday, November 15, 2011 - link
I was waiting for X79 but after I read the initial reviews I bought a Z68 motherboard and 2500K cpu for gaming.C300fans - Tuesday, November 15, 2011 - link
Me too. 999$+X79 for 0% improvement in gaming. What a crab! Bulludozer seems not that crab comparing to 3960x overall.yankeeDDL - Tuesday, November 15, 2011 - link
Making unsubstantiated claims about something that is non-intuitive falls, in my dictionary, under fanboy-ism (if that's a word).The fact that Win7 "runs better" on a certain, relatively old, PC, is one thing. Stating that Windows7 is faster than XP (in spite of a documented benchmark proving otherwise) is another one.
Like I said, you can compare OS in terms of HW support, ease of use, even responsiveness, however, neither of those translate into one OS beinf "faster".
Faster means that when you run a benchmark (pick any of the ones that Anand run in this article), you get a noticeable increase in speed.
The OSes provide the infrastructure to run applications, they cannot provide any fundamental speed difference, unless, of course, you have a PC without enough RAM, for example, and in that case the OS that uses less RAM will have an obvious advantage (because it offers more "free" RAM for apps to run), but that again, has nothing to do with one OS being faster: if anything, it is more efficient.
I have 4GB on both my laptop (Win7) and on my desktop (WinXP) and the difference is negligible: I nearly always have more than 2GB of RAM committed, so it is no surprise that on your PC Win7 with ReadyBoost is faster: just spend ~$15 on 2GB of RAM and you'll see a huge performance improvement both on XP and 7.
jmelgaard - Tuesday, November 15, 2011 - link
So "Faster" must not apply to the OS's capability to respond to the user, it must only apply to the OS's capability to server application requests?...Wait what?...
Kob - Tuesday, November 15, 2011 - link
You guys need to look at the engineering of your requests: 6 sata3 ports require feeding 6*6gb/s = 36 Gb/s data, while the total max theoretical mem bw of the chipset is 37 Gb/s. Can't do that while also taking care of OS, apps and video memory requirements.cbutters - Monday, December 12, 2011 - link
6*6gb/s isn't going to be happening constantly.....you build out one bridge that has a certain amount of bandwidth, 12GB perhaps, I don't know, and let the ports use the available shared bandwidth, doesn't mean you can't add additional ports, this is one of the benefits of serial interfaces.C300fans - Tuesday, November 15, 2011 - link
Intel Gulftown 6C 32nm 6 1.17B 240mm2Intel Sandy Bridge E (6C) 32nm 6 2.27B 435mm2
SB-E, What a crab! Double Transistors, Double size, merely 20% gain from SB 2600k. 999$ for this? I would rather get 2 pcs Interlagos 6200 instead.
sna1970 - Tuesday, November 15, 2011 - link
using 5870 CF to show us that dual 8x PCIE are same as dula 16x is a mistake I am shocked some one like you fall in ...you should have tested 6990 in CF , or 590 ... and see the difference between 16x SLI/CF and 8x SLI/CF
and how do you consider a 5870 a MODERN GPU ?
Quote : "Modern GPUs don't lose much performance in games, even at high quality settings, when going from a x16 to a x8 slot."
Answer : WRONG . try high end dual GPU cards in SLI/CF !