MAXXMEM Benchmark Results Thread

[Chicken Patty]

Hello all, welcome to the MAXXMEM Benchmark Results Thread. I made this thread so that we can all post our scores up and have a decent idea of what to expect from our RAM and just to have it serve as some sort of database fro those looking for overclocking info on certain RAM kits or what not.

If you are going to post scores, please post them in the below format. The MAXXMEM window must in the post along with the below format.

User Name-
Intel/AMD-
RAM-
RAM Voltage (Optional)-
CPU-
BOARD-
OPERATING SYSTEM-
RAM TIMINGS-
RAM SPEED-
MEM COPY (MB/sec)-
MEM READ (MB/sec)-
MEM WRITE (MB/sec)-
MEM LATENCY (ns)-

Please feel free to chip in your own results if you want to

YOU CAN DOWNLOAD MAXXMEM HERE
Stats.
CLICK HERE


I'll start it off:

[cadaveca]

If you want to make this a real good research effort, you should start @ 1066 mem, and 1600mhz NB.

First, compare how increase in NB multi affects bandwidth and L3.

Then, go to 1333, and do the same, however, due to AMD's own recommendations, NB speed should be 2000mhz minimum, but I think it might be beneficial to see from 1600NB up.

Then, of course, 1600mhz mem. Again, AMD recommends 2400NB AT LEAST, but again, it would be interesting to see the scaling...

I'd also be interested in how CPU multi affects that too. For example, for 2.8ghz...then 3.2ghz, 3.6ghz, 4.0ghz...

Of course HWBOT has an article up kinda doing the same thing...and I've also done such testing myself...but another source of info would be great...

[Chicken Patty]

Originally Posted by cadaveca

If you want to make this a real good research effort, you should start @ 1066 mem, and 1600mhz NB.

First, compare how increase in NB multi affects bandwidth and L3.

Then, go to 1333, and do the same, however, due to AMD's own recommendations, NB speed should be 2000mhz minimum, but I think it might be beneficial to see from 1600NB up.

Then, of course, 1600mhz mem. Again, AMD recommends 2400NB AT LEAST, but again, it would be interesting to see the scaling...

For now what I wanted to do since rig is up and running already at these settings is try and get better from here. once I found my best settings then I'll go ahead and drop the NB speed and loosen up timings and stuff and go from there. Kinda backwards but yeah

[cadaveca]

OK, so you just want to optimize things for yourself, then?

Run 1600mhz, 2600NB, best timings the ram can do, and call it done.

Also, grab the app Maxxmem, as another testing tool.

[Chicken Patty]

Originally Posted by cadaveca

OK, so you just want to optimize things for yourself, then?

Run 1600mhz, 2600NB, best timings the ram can do, and call it done.

At first. Then I'll go ahead and do as you stated above.

[cadaveca]

Might be interesting to see how different boards deal with it too...maybe some of us here can toss up some numebrs with you, and we might see what the best board is for memory?

Mind you, my ideas are kinda alot of work...

[Chicken Patty]

Originally Posted by cadaveca

Might be interesting to see how different boards deal with it too...maybe some of us here can toss up some numebrs with you, and we might see what the best board is for memory?

Mind you, my ideas are kinda alot of work...

That's fine with me. What I would appreaciate is that if anybody does this. Run a few benches and compile them into one PM. Then I can go ahead and add to the OP. I will hopefully be receiving my X6 and Crosshair IV this week. If I do this is the RAM i'll be using anyways. How's that for another board?

[Chicken Patty]

Guys, OP is updated with results from another member. It was his laptop but as satted, this is just for comparison. Why not?

[Chicken Patty]

Ok, I haven't restarted my PC so I haven't lowered my CAS yet, but I did lower a few timings and my memory latency dropped a bit but my L3 Cache went up a bit, any input on that?
8-8-8-20

8-7-7-20



Ran it again and it seems to vary. Overall this one is better. My Memory is .1 higher, but my L3 is .4 lower.

[JrRacinFan]

Here's where I have mine set to CP.



You should be able to tighten timings some.

EDIT:

The fluctuation was from your CPU PLL Clockgen, look at your HTT speed. Add a notch to that in bios from 1.2v to 1.3v and it should do it less.

[mastrdrver]

CP, I'll send you a PM with the settings I've tried.

Originally Posted by cadaveca

OK, so you just want to optimize things for yourself, then?

Run 1600mhz, 2600NB, best timings the ram can do, and call it done.

Also, grab the app Maxxmem, as another testing tool.

Agreed on Maxxmen. You'll be able to tell if any timings you played with are going to show any tangible results in benches. Increases in that bench (and by how much of one) have always proved to show increases in other benches.

Also, fwiw CP I found that with 1333 7-7-7-18 1T ram and resulting 2600 cpu-nb that I needed 3600 on the core to fully benefit from it all. This was on the same Gigabyte board with F3l bios and a C3 955. F7 actually showed better memory performance and that board also takes a big memory performance hit when at least Raid 0 is enabled.

I know I've seen it mentioned on XS cadaveca, and I've just recently been able to verify this some what, but compared to my Gigabyte 790fxt-ud5p with the same ram and my 555 C3 that fully unlocks without problems, clock for clock my Gigabyte board is faster even though I've run up the HT Link to 2600 on the MSI board. Though, it is a 790GX so I'm not sure what the implications are compared to the 790FX. I would think it should be pretty similar given similar hardware and the same ram sticks.

[Chicken Patty]

Originally Posted by mastrdrver

CP, I'll send you a PM with the settings I've tried.



Agreed on Maxxmen. You'll be able to tell if any timings you played with are going to show any tangible results in benches. Increases in that bench (and by how much of one) have always proved to show increases in other benches.

Also, fwiw CP I found that with 1333 7-7-7-18 1T ram and resulting 2600 cpu-nb that I needed 3600 on the core to fully benefit from it all. This was on the same Gigabyte board with F3l bios and a C3 955. F7 actually showed better memory performance and that board also takes a big memory performance hit when at least Raid 0 is enabled.

I know I've seen it mentioned on XS cadaveca, and I've just recently been able to verify this some what, but compared to my Gigabyte 790fxt-ud5p with the same ram and my 555 C3 that fully unlocks without problems, clock for clock my Gigabyte board is faster even though I've run up the HT Link to 2600 on the MSI board. Though, it is a 790GX so I'm not sure what the implications are compared to the 790FX. I would think it should be pretty similar given similar hardware and the same ram sticks.

I'll look into maxxmem when I get home today and give that a shot. Thanks

[Super Sarge]

Here is my Bench Mark

[overclocking101]

have you tried using memset?? I think it works on AMD rigs im not sure if not try cpu tweaker. they allow on the fly tming changes the only one thats locked is the CAS the very first timing the rest are fully customizable.

[Chicken Patty]

memset didn't want to work, CPU tweaker did though. Guys hold off on the runs, Imma be changing it up a bit, just need a day or two till I find more time. I mean I don't mind you posting, but I'm not adding anything else to the OP for now.

[Chicken Patty]

Ok, OP updated. Let's get the party rollin'

Google docs spreadsheet linked to OP to keep track of things.

[Paulieg]

Originally Posted by Chicken Patty

That's fine with me. What I would appreaciate is that if anybody does this. Run a few benches and compile them into one PM. Then I can go ahead and add to the OP. I will hopefully be receiving my X6 and Crosshair IV this week. If I do this is the RAM i'll be using anyways. How's that for another board?

The Crosshair IV is on it way to you today. It handles DDR3 better than any other AMD board that I've owned.

[b2vFAh6J]

Hi guys, here's mine.

My bios settings are definately not optimized. I'm thinking about creating a small, micro-xp installation with some small benchmark/cpu burn utilities.

[Chicken Patty]

Originally Posted by Paulieg

The Crosshair IV is on it way to you today. It handles DDR3 better than any other AMD board that I've owned.

thanks bro, really looking forward to that

Originally Posted by b2vFAh6J

Hi guys, here's mine.
http://img683.imageshack.us/img683/4...xmem271947.jpg
My bios settings are definately not optimized. I'm thinking about creating a small, micro-xp installation with some small benchmark/cpu burn utilities.

Thanks for posting your results. I'll update the spreadsheet when I get home

[JrRacinFan]

User Name-JrRacinFan
Intel/AMD-Intel
RAM-2x2GB Nanya 1066 CL7
CPU-Core i3 530
BOARD-Asus P7P55 WS
OPERATING SYSTEM-Windows 7 x64 Ultimate
RAM TIMINGS-7-7-7-20 1T
RAM SPEED-628/1256
MEM COPY (MB/sec)-8670
MEM READ (MB/sec)-8255
MEM WRITE (MB/sec)-6985
MEM LATENCY (ns)-84.7

[mlee49]

Just what I've been waiting for, another Memory benchmark!

Triple Channel FTW!!!!

Small overclock on the ram, fairly tight timings too(for 1600Mhz):



User Name- mlee49
Intel/AMD- Intel
RAM- Corsair Dominator GT(1600Mhz)
CPU-i7 920
BOARD-Evga X58(E758)
OPERATING SYSTEM-Win7 64x
RAM TIMINGS-7-7-7-21
RAM SPEED-1624Mhz
MEM COPY (MB/sec)-16317MB/s
MEM READ (MB/sec)-14348MB/s
MEM WRITE (MB/sec)-13065MB/s
MEM LATENCY (ns)-50.7ns

[Yukikaze]

Here is the benchie for the pair of systems I have next to me here:

User Name-Yukikaze
Intel/AMD-Intel
RAM-12GB (6x2GB) OCZ Platinum DDR3 1600Mhz (Stock: 7-7-7-24, 1.65v)
CPU-i7 975
BOARD-DFI LanParty DK X58-T3eH6
OPERATING SYSTEM-Windows Vista 64-bit
RAM TIMINGS-6-6-6-15-1T
RAM SPEED-1333Mhz
MEM COPY (MB/sec)-16663MB/s
MEM READ (MB/sec)-15090MB/s
MEM WRITE (MB/sec)-13181MB/s
MEM LATENCY (ns)-50.1ns



User Name-Yukikaze
Intel/AMD-AMD
RAM-2GB (2x1GB) OCZ Platinum EL DDR1 400Mhz + 1GB Geil DDR1 400Mhz
CPU-Opteron 180
BOARD-Asus A8N32-SLI Deluxe
OPERATING SYSTEM-Windows Vista 64-bit
RAM TIMINGS-2-3-3-6-2T
RAM SPEED-400Mhz
MEM COPY (MB/sec)-3799MB/s
MEM READ (MB/sec)-5820MB/s
MEM WRITE (MB/sec)-3865MB/s
MEM LATENCY (ns)-70.7ns



What's weird, though, is that the Everest reading I have for the Opteron 180 is way different, with much faster copy and much lower latency. I do not get a change like this on my i7, though:

[Chicken Patty]

Thanks guys, hopefullythis thread gets as big as my wprime thread. It's been fun over there.

For the ones who posted without the format in the OP. Please edit your post with the format filled out so I can add your scores when I get home. Thanks.

[AthlonX2]

AthlonX2
Intel
GSkill Trident
Xeon W3520
eVGA E758
Windows 7 x64
99924
2000Mhz
MEM COPY (MB/sec)-21857
MEM READ (MB/sec)-18936
MEM WRITE (MB/sec)-17104
MEM LATENCY (ns)-41.2

[somebody]

Any particular reason to use 1.50

Latest is 1.84

# v1.73, MaxxPI.net / HWbot.org submit added (30/12/2009)
# v1.74, minor bugfixes (03/01/2010)
# v1.80, maintenance update, Quick compare added (19/04/2010)
# v1.82, minor bugfixes at Quick compare (22/04/2010)
# v1.84, several bugfixes and other issues (14/06/2010) NEW!

http://www.maxxpi.net/pages/download...---preview.php