Memory Scaling on Haswell CPU, IGP and dGPU: DDR3-1333 to DDR3-3000 Tested with G.Skill
by Ian Cutress on September 26, 2013 4:00 PM ESTThe activity cited most often for improved memory speeds is IGP gaming, and as shown in both of our tests of Crystalwell (4950HQ in CRB, 4750HQ in Clevo W740SU), Intel’s version of Haswell with the 128MB of L4 cache, having big and fast memory seems to help in almost all scenarios, especially when there is access to more and more compute units. In order to pinpoint where exactly the memory helps, we are reporting both average and minimum frame rates from the benchmarks, using the latest Intel drivers available. All benchmarks are also run at 1360x768 due to monitor limitations (and makes more relevant frame rate numbers).
Bioshock Infinite: Average FPS
Average frame rate numbers for Bioshock Infinite puts a distinct well on anything 1333 MHz. Move up to 1600 gives a healthy 4-6% boost, and then again to 1866 for a few more percent. After that point the benefits tend to flatten out, but a bump up again after 2800 MHz might not be cost effective, especially using IGP.
Bioshock Infinite: Minimum FPS
Unfortunately, minimum frame rates for Bioshock Infinite are a little over the place – we see this in both of our dGPU tests, suggesting more an issue with the title itself than the hardware.
Tomb Raider: Average FPS
Similar to Bioshock Infinite, there is a distinct well at 1333 MHz memory. Moving to 1866 MHz makes the problem go away, but as the MHz rises we get another noticeable bump over 2800 MHz.
Tomb Raider: Minimum FPS
The minimum FPS rates shows that hole at 1333 MHz still, but everything over 1866 MHz gets away from it.
Sleeping Dogs: Average FPS
Sleeping Dogs seems to love memory – 1333 MHz is a dud but 2133 MHz is the real sweet spot (but 1866 MHz still does well). CL seems to make no difference, and after 2133 MHz the numbers take a small dive, but back up by 2933 again.
Sleeping Dogs: Minimum FPS
Like the average frame rates, it seems that 1333 MHz is a bust, 1866 MHz+ does the business, and 2133 MHz is the sweet spot.
Facebook
Twitter
RSS
89 Comments
View All Comments
Rainman11 - Tuesday, October 1, 2013 - link
The gaming segment was utterly pointless. Show the difference using a resolution of at least 1080p or don't even bother including it.Anonymuze - Tuesday, October 1, 2013 - link
I'm really curious to see a similar test on HD5000 or (28W) HD5100 - they don't have the benefit of EDRAM like the HD5200 and should be much closer to being memory bandwidth limited than HD4600.Anonymuze - Tuesday, October 1, 2013 - link
..."should be much closer to being memory bandwidth limited"I meant to say "should be much closer to memory bandwidth limits" or "should be much more memory bandwidth limited" - pick one :P
Kathrine647 - Wednesday, October 2, 2013 - link
like Gregory said I am alarmed that a stay at home mom able to earn $5886 in 1 month on the internet. visit their website............B u z z 5 5 . com open the link without spacesHrel - Thursday, October 3, 2013 - link
This is a lot of pages on content that all just tells you to buy 1866-CL9. Good to know.SetiroN - Friday, October 4, 2013 - link
Ian, you REALLY should include code compilation benchmarks.80% of the people I know who actually need a powerful CPU/RAM/SSD combination use it to build software.
You took the time to test IGP performance (who the spends money on RAM to play on an HD4000?) when you could have provided much more useful data. :)
dreamer77dd - Saturday, October 5, 2013 - link
AMD might like higher speed RAM then Intel. That could be interesting article also.Laststop311 - Sunday, October 6, 2013 - link
This article just confirmed my suspicions, that this more expensive faster ram basically has no effect on your system. Basically anything 1866+ is going to be relatively the same performance. I use 2133mhz CAS 8 ram in my system and am totally happy and only paid 105 for 4x4GB kit.SmokingCrop - Sunday, October 27, 2013 - link
What a useless test.. Now we don't even know if resolution matters..No one is going to be doing crossfire so (s)he can play on 1 monitor with 1360x768 pixels..
qiplayer - Saturday, November 2, 2013 - link
I don't understand testing a 3000mhz kit and to evaluate gaming performance use that resolution (extremely low) and even not one gpu.I would suggest to once test the difference with this very interesting test on a triple hd resolution with 2 or 3 gpu. Or even better, as we talk about memory for the enthusiasts, cpu should be overclocked, gpu should be at least 2 and overclocked.
Te title cud be: Aiming at 120hz on 5800x1080, how much to spend on the ram?
Maybe it comes out that 150$ more on memory are enough for 5% higher fps, that are not nothing when spending already some $$$$ on gpu to get the best, another $$$ on cpu and $$$$to put all on water.