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
yankeeDDL - Monday, November 14, 2011 - link
That's what I was doing: blaming AMD. Intel is doing what any company that is interested in making money/profit would do.yankeeDDL - Monday, November 14, 2011 - link
Sorry ... I replied to the wrong post :) I meant to hit the one below!JlHADJOE - Tuesday, November 15, 2011 - link
If you compare it to the extreme edition chip, then Bulldozer looks like good value. But then the 3960X is a halo model for those people who care nothing about price.Considering the 3930K gives you 95% the performance of the 3960X for 50% of the price (see xbitlabs), there's really no reason to get the X-edition chip unless you are building a system purely for bragging rights.
Now when compared to the 3930K, the FX8150 doesn't look nearly as good. If we consider total platform costs with either system having a $300 motherboard and $200 in ram, then we're looking at something in the region of $750 for BD, vs $1000 for the SB-E. +$250 is small change for double the performance at a similar power envelope.
yankeeDDL - Tuesday, November 15, 2011 - link
JIHADJOE, yes, that was my point.The 3960X is -arguably- the fastest CPU available, but it is faster by a tiny margin, while being radically more expensive than anything else.
So yes, nothing looks as bad in terms of price/performance ratio, not even the FX8150. And that's, basically, bad for everyone (except Intel)
actionjksn - Monday, November 14, 2011 - link
I agree too AMD's poor Bulldozer performance is having a huge effect on what we can get from Intel and at what price. And I blame AMD not Intel, because Intel or any other company is supposed to do what's best for their company. Heck if Intel did what we want they would probably cause serious harm to AMD. Because it would make AMD even less competitive. And I don't think Intel really wants to put AMD out of business.yankeeDDL - Monday, November 14, 2011 - link
That's what I was doing: blaming AMD. Intel is doing what any company that is interested in making money/profit would do.GeorgeH - Monday, November 14, 2011 - link
One of the bigger advantages of this platform to me is the 8 DIMM slots. However it was rumored that the first revision of SB-E had/has VT-D problems, which spoils things a little bit as VMs are one of the bigger reasons for lots of RAM. Can you confirm or deny if there are VT-D issues?Anand Lal Shimpi - Monday, November 14, 2011 - link
VT-d is supported, checking to see if there are any functional issues now.Take care,
Anand
GeorgeH - Monday, November 14, 2011 - link
ArsTechnica is reporting that VT-D is broken, but they don't cite any sources. A short article explaining what VT-D is for those who don't know and what (if anything) is broken might be in order.Filiprino - Monday, November 14, 2011 - link
That thing is really big!As for Quick Sync, it's not really useful. If you want quality you'll have to use x264, and with lower qualities x264 has some presets that are near as fast as Quick Sync.
The winner combo is LGA2011+Kepler/Souther Islands.
If you have a hole in your pocket you can throw in a dual socket motherboard, some liquid cooling and a big SSD.