Memory Latency Impact on Performance
We just looked at the impact of memory bandwidth on performance, but what about latency? Let's first by adjusting the CAS latency from our default of 2 clocks up to 3 clocks. Almost all DDR400 these days is CAS 2 memory, but older memory may have a higher CAS latency or you may have to increase your CAS latency when overclocking to gain more memory bandwidth, so what kind of a performance hit is there when going from CAS 2 to CAS 3?
at_canals_08 |
at_coast_05 |
at_coast_12 |
at_prison_05 |
at_c17_12 |
|
Tcl = 2 |
116.12 |
140.43 |
123.37 |
113.69 |
83.15 |
Tcl = 3 |
115.52 |
137.07 |
121.91 |
113.37 |
79.92 |
At worst, CAS 2 memory seems to be about 5% faster than CAS 3 memory when looking at at_c17_12, our most CPU intensive test. While 5% alone isn't anything major, combine that with a number of other performance tweaks and they can definitely begin to add up.
Now let's look at keeping Tcl (CAS latency) fixed at 2 clocks, but vary Trcd timings from 3 up to 6 clocks:
at_canals_08 |
at_coast_05 |
at_coast_12 |
at_prison_05 |
at_c17_12 |
|
Trcd = 2 |
116.12 |
140.43 |
123.37 |
113.69 |
83.15 |
Trcd = 3 |
115.71 |
136.99 |
122.46 |
113.08 |
79.97 |
Trcd = 4 |
113.92 |
134.42 |
120.87 |
112.38 |
79.83 |
Trcd = 5 |
113.42 |
131.82 |
119.34 |
114.79 |
79.12 |
Trcd = 6 |
113.23 |
128.26 |
117.56 |
111.15 |
77.4 |
For the most part we saw no real changes when adjusting Trcd, the one exception being at_coast_05 which actually showed a pretty big difference between a Trcd value of 2 and higher latency values.
Next we'll look at adjusting Trp:
at_canals_08 |
at_coast_05 |
at_coast_12 |
at_prison_05 |
at_c17_12 |
|
Trp = 2 |
116.12 |
140.43 |
123.37 |
113.69 |
83.15 |
Trp = 3 |
115.6 |
139.24 |
123.13 |
116.35 |
82.09 |
Trp = 4 |
115.85 |
138.88 |
122.98 |
113.16 |
82.05 |
Trp = 5 |
114.84 |
138 |
122.65 |
112 |
80.98 |
Trp = 6 |
114.5 |
136.95 |
121.96 |
115.61 |
80.95 |
Here we see very little impact on performance.
Putting them all together we can see what the overall impact on using fast DDR400, higher latency DDR400 and extremely high latency DDR400 will be:
at_canals_08 |
at_coast_05 |
at_coast_12 |
at_prison_05 |
at_c17_12 |
|
2-2-2-10 |
116.12 |
140.43 |
123.37 |
113.69 |
83.15 |
3-3-3-10 |
114.47 |
134.11 |
120.64 |
112.62 |
80.56 |
3-6-6-10 |
110.74 |
123.76 |
114.75 |
112.17 |
73.8 |
Our standard 2-2-2-10 memory does actually offer reasonable performance benefits in Half Life 2 compared to DDR400 with higher timings such as 3-3-3-10 or the unrealistically high 3-6-6-10.
First and foremost Half Life 2 does appear to be rather dependent on memory bandwidth, but it is also quite appreciative of low latency memory as well. If you're wondering whether being able to run memory at low timings and high clock speeds is important, when it comes to Half Life 2 performance it is.
Facebook
Twitter
RSS
68 Comments
View All Comments
zhangping0233 - Thursday, January 5, 2012 - link
Nice job, If you need any flashlight, pls contact Xecconlight.com feel free.PrinceGaz - Friday, January 28, 2005 - link
by the way, you can't mod a 9500Pro to a 9700Pro, the 9500Pro circuit-board only has a 128-bit memory-bus and there's no way you can change it to 256-bit.PrinceGaz - Friday, January 28, 2005 - link
#55- the cut-off point where you can't really tell the difference of a higher framerate, is when the framerate exceeds the monitor refresh rate.If your monitor is updating the display 85 times per second (a common setting for cheaper CRT displays), then a *minimum* framerate of higher than 85fps makes no difference. With flat-panels, refresh-rates of 60-75hz are more common so you don't even need to maintain 85fps. A faster graphics-card is still worthwhile though as it allows you to crank up AA and Aniso settings (8x anti-aliasing is lovely).
maestroH - Friday, January 28, 2005 - link
Coming from the recent dark age of a P4 1.7/9500Pro(@9700Pro) combi, this is great article to decide on my new machine. Although HL2 is playing quite nicely, I am lucky to have a one-off opportunity to buy myself a FX-55/X800XT combi.Never having experienced even any fps close to what's on these charts, something in the back of my mind keeps saying that a 10-20 fps more when you are already over 100 fps, will make no difference to the experience except a bigger hole in my wallet. Can anyone tell me where the 'cut-off' point is where even the most discerning of gamers cannot see/feel the difference? Knowing that dual core is coming up (even though games for them still need to be made), would buying a 3500+ be smarter or should I go for the FX-55 simply because I can (only this once)? Thx Anand for the article.
essjae - Thursday, January 27, 2005 - link
Those graphs look nice, but they don't really mean much. Based on similar graphs and results I just bought a a64-4000+ and MSI Neo2 Platinum to replace my P4-3.2GHz and Asus P4c-800E.With the same ATI X800XT Platinum, memory, and hard drives, I can't see any difference between then, in fact, the P4 seemed to play smoother.
Do I have any proof, no, other than playing half-life 2 on my p4 was more enjoyable.
Spacecomber - Thursday, January 27, 2005 - link
Can anyone offer some insight into why the extreme edition northwoods did as poorly as they did? The 3.2GHz EE could barely keep up with the 3.0GHz Prescott; so, it's more than the raw clock speed of the high end Prescotts.Could it be related to running the Northwood on a platform really intended for Prescotts?
Space
mixpix - Thursday, January 27, 2005 - link
Awesome article. It was exactly what I've been looking for. My 2600+ is not cutting it with my 6600GT AGP and I was thinking it was the CPU that was limiting preformance.TheCanuck - Thursday, January 27, 2005 - link
Firing Squad did a review on the Athlon XP performance with HL2 a while ago:http://www.firingsquad.com/hardware/half-life_2_at...
The 3200+ XP got about 95fps in the Canals_09 demo at 1280x1024 with an X800 XT PE. Not sure how well it compares to the Canals_08 that Anandtech uses, but I doubt the difference would be that great.
Guspaz - Thursday, January 27, 2005 - link
I'm very dissapointed with this article. I have been eagerly awaiting it for ages, expecting to see how Half-Life 2 scales down to lower speed processors; I've long maintained that low end processors like an AthlonXP 1900+, run the game quite poorly.The big deal with HL2 is that it was supposed to run on much older computers. But nobody seems to have benchmarked it on anything but pretty new hardware. I expected that a CPU scaling article would cover that, in fact I thought that was the entire point.
Visual - Thursday, January 27, 2005 - link
i'd be curious to see normal 6600, as well as maybe some lower-end ati card in the comparison :)ok ok, i know this isn't a GPU shootout, but still...