• Welcome to Overclockers Forums! Join us to reply in threads, receive reduced ads, and to customize your site experience!

Should I be concerned about this Sandra Benchmark?

Overclockers is supported by our readers. When you click a link to make a purchase, we may earn a commission. Learn More.

U235

New Member
Joined
Dec 8, 2008
Long story short, I used to be really into overclocking and ran a business selling overclocked PCs, but 6 years ago I went to work for Dell and sort of fell out of overclocking. Since then I've worked for a company that makes relatively mobile peripherals and I nearly gave up using desktops altogether. Anyways, I finally decided to build up a desktop since I had access to many free parts.

I scraped together the following system (only purchase being the CPU, everything else was free due to co-marketing things I've done with the companies involved over time). Thus, please don't criticize the components, they weren't specifically selected (except CPU):

CPU: Intel C2D E8400 Wolfdale
Cooler: Coolermaster TX2
Mobo: MSI K9 Neo-F (DDR2)
RAM: Buffalo Firestick 1066 (2x1GB) - fsx1066d2c
PSU: Enermax EG565P-VE

Anyways, my current setup/config is as follows:
468x9 = 4211 MHz
FSB = 1871 MHz
DDR2 = 936 MHz
Timings = 4-4-4-12 (in bios)
I've ran higher FSBs but would get the occasional issue or two (mainly had a trouble soft-rebooting).

Here are some pics:

bc16d5450cf04d45ac2783df993b2f34.jpg

779fb15ca132be01f54ca4451580efc8.jpg


Now the problem is with Sandra. I know it's synthetic and all, but I've never seen numbers this off before. My memory benchmarks SUCK compared to the reference profiles:

08ad23a8bd8e33c5930398241750fe8c.jpg


At the bottom it says my timings are bad (which I don't think they are considering the FSB) and that the Large Memory Pages can't be used (I went ahead and supposedly fixed that in Vista by adding my account to the group policy, but that hasn't made the error go away).

Should I just ignore these or is their truth in these lousy numbers?

Thanks,
U235
 
Grr, so much for that.

The CpuZ shots just back up what I said, and I assure you they're true.

Here's the Sandra (click to enlarge):
 
umm you comparing to computers that have the ram running faster then yours.... cpuz shows the rams SDR rate, which is 467.9x2 for DDR2 rate equals DDR2-935, DDR2-800 4-4-4-12 would be faster then your current 4-5-5-12. even the 570i setup with DDR2-1066 is going to be faster even with cas5 timings.

i dont see a problem, have you run 3dmark06?
 
umm you comparing to computers that have the ram running faster then yours.... cpuz shows the rams SDR rate, which is 467.9x2 for DDR2 rate equals DDR2-935, DDR2-800 4-4-4-12 would be faster then your current 4-5-5-12. even the 570i setup with DDR2-1066 is going to be faster even with cas5 timings.

i dont see a problem, have you run 3dmark06?

Even to that discrepancy? I had reference markers in there with PC5300 and it was about the same as mine (a little faster in fact).
 
well i all i get from that bench is a bargraph, try using everest and see what you get then...
 
+1 on everest...

SS seems to of been going downhill on the benchmarks...I dont even use SS anymore...especially since they put the advertisment links on the free versions...
 
It looks a tad low, nothing serious, maybe your cheapo mainboard is shaving off a few hundred megabytes. Theres only so much one can do for bandwidth on the antique Intel FSB/MCH, even DDR3-1800 will not score much above 7000 integer with a core2duo. OC'ed Core i7 will score well above 20k integer with dual channel and around 30k in tripple channel with DDR3-1800 so thats the way to go if you need bandwidth. But who cares, it doesnt mean much in real life performance anyway.
 
that must not be the correct model board, all results i see with some slight deviation. pull up am2 boards, based on the NV-5xx chipset.
 
He must have gotten part of the mainboard model right, all Neo-F are close to the bottom of the line, or most MSI boards are imho. I guess P35 Neo-F.
 
He must have gotten part of the mainboard model right, all Neo-F are close to the bottom of the line, or most MSI boards are imho. I guess P35 Neo-F.

It's a P45 Neo-F. I'm upgrading some an AMD X2 5000+ with PC5300. I found an AM2 X2 4800+ with PC5300 reference on there and it was faster than my Sandra score.

I'm possibly concerned about the warning message on the Sandra bench about locked page memory which I've yet to get go away.
 
I found an AM2 X2 4800+ with PC5300 reference on there and it was faster than my Sandra score.
No surprise there, AMD has a superior memory controller.
I'm possibly concerned about the warning message on the Sandra bench about locked page memory which I've yet to get go away.
Did you do any tweaks to Vista? Maybe remove the /3gb switch from boot.ini if its there.
 
yea you cant compare X2 memory scores to intel core 2. since the X2 has memory controller on die vs in the northbridge for the core 2. even i7 has a IMC like the X2, plus i7 has tri-channel ram so you cant compare there either.
 
Thanks guys, how do you feel about this comparison. I'm concerned that I should be above the PC2-6400 P35 example...

PS - This is at ~150MHz slower speed due to a cold boot problem I was having, so this is @ DDR2-912

Evilsizer, nice to see a fellow Austinite in here! Thanks for all the help!

 
Last edited:
well this might be dumb to ask but sometimes its the simpliest things that get over looked. have you installed the chipset drivers for the mobo? get them directly from intel, not the manufacture of the board. in either case though at ddr2-912 cas5, you should be seeing higher bandwidth numbers. have you tried using everest yet?
http://www.lavalys.com/

yep,np man! :soda:
 
Thanks guys, how do you feel about this comparison. I'm concerned that I should be above the PC2-6400 P35 example...
Thats a quadcore, they score a little higher in these benchmarks. Considering that you have a nice overclock on your CPU you should have somewhat higher score but i still think its close enough, i wouldnt worry.
 
Thanks for the advice guys, I felt they were a little slow when comparing to references.

Evil - I've installed the latest chipset driver from MSI (not Intel). I'll try Intel's, but I'm out of town until Monday and won't be able to do an A-B test until then, but I'll be sure to report my results. I wanted to try Everest, but for some reason I got the impression that they no longer offer a free version/demo. I just double-checked and for some reason all I see is "Purchase" links. I'm not sure I'm willing to purchase software to pursue this at this exact moment; may want to do more troubleshooting first. Am I missing a free version somewhere?

Thanks to all for the help!
 
Back