PC performance not as good as expected

yeah, to be honest I'd recommend using other benchmarking software (3d mark etc) as they give a more accurate rating of how your computer would perform in games.
 
It just whats left on 1 drive, i have 1.5tb

neo-omega said:
yeah, to be honest I'd recommend using other benchmarking software (3d mark etc) as they give a more accurate rating of how your computer would perform in games.
Click to expand...


No it certainly will not. There is no ai, and many other things lacking in 3d mark to give any real gaming performance query results.
 
Thanks everyone, I ran a 3D Mark test:

n2fwhd.jpg


For some reason my Asus Commando won't allow me to change the CPU Frequency in the BIOS.

2j1q4pw.jpg


It lets me set the AI Tuning to manual but when I press enter on the CPU frequency itself, nothing happens... I'm probably not doing something rather simple, but can anyone point me in the right direction? :)
 
twist3d0n3 said:
tis pretty poo memory by todays standards...

my pc-8500 1066MHz ram only hits 6.6
Click to expand...

Well to throw a spanner in that works... My windows 7 Memory score is 7.2 with OCZ DDR2 PC-5400, I have 2x1GB and 2x2GB (total 6) in matching pairs, and correctly installed for dual channel operation.

The actual clock speed of the memory on Core 2 Duo's and quads is a lot less important than with the newer i7's as faster memory often exceeds the available bandwidth on the cpu's FSB, and is therefore delayed in wait cycles.

My full windows 7 scores are CPU 7.2, Memory 7.2, Graphics 6.0, Gaming Graphics 6.0, and Hard disk 5.9.

Q6700 at 2.96Ghz, 6GB Ram @ 296mhz(592DDR), and a Geforce 8800GTX that artifacts at anything more than stock speed :(.

Before you ask about the 2.96Ghz.. let me answer, I cant be bothered to solder the debug jumper on my ancient intel motherboard, so I am stuck at stock volts for my Q6700 :/, my G0 has a very low VID so the motherboard isnt giving the chip much power to play with :P. But still my 8800GTX is the biggest bottleneck, so I dont really care much right now.
 
Last edited:
Have you tried pressing + or - on the CPU frequency, a lot of motherboards use that system to change the clock speed without the need for pressing enter.

PS, if you do change the clock speed of the CPU the ram will be bumped up by the same percentage, potentially pushing it above its rated value. I normally overclock the cpu first while temporarily setting the memory to its slowest setting (Normally 266mhz/533DDR/PC4200). As you overclock the actual speed of the ram increases to match the CPU in bandwidth at this setting, and has the best odds of tight timings when your finally done.
 
Last edited:
Ah ha, you are correct. How daft do I feel?

Ahh cool, I'll drop the dram frequency down too. I'm gonna use SP2004 Orthos to stress test. I'll assume that's ok for now. Thanks!!

edit: Should I run 4 copies of this then all on each core??
 
Last edited:
If the test only uses a single core, then yes, by all means run up 4 copies, and if you really want to be sure use the task manager to set affinity for each copy and lock each one to a specific core, to prevent windows from moving them from core to core (which will cause slightly less overall load each time a copy moves from core 0/1 to core 2/3 as the shared cache on Q6xxx and Q9xxx are split between each pair of cores not all 4 as a whole.

You can step up the memory speed again later if you really want to, but in my experience Core 2 Duo and Quads do not benifit from ram that is clocked in excess of the fsb anyway, unless for some reason dual channel is broken, in which case clock away on the ram too :).

Strictly speaking, single channel @ 1066 on a stock speed E6600 is probably slightly faster than dual channel @ 533 due to latency and raw speed on each memory location. In practice because Intel's branch predictor and cache loaders are so efficient, memory performance is generally a secondary area to improve performance.
 
Forgot to mention, as your using 4 sticks of ram (as am I), you may find you need additional voltage to keep the memory stable at higher speeds, Corsair should be fine with at least 1.9V, some versions actually need more though, Corsair are pretty good company, they normally print the recommended voltage on the label, so if they print 2.1V on there somewhere, A) it will need 2.1V for the highest speeds, and B) they will honor the warrenty at that voltage.

That said, starting at DDR@266mhz (PC4200) the odds are the ram will be error free and the CPU will high its wall first :).
 
Woo, excellent stuff. I started a CPU test just to see how it ran on my default settings and on 4 instances of the program (one core assigned to each) it froze up. But on 2 instances of the program it went fine for 20 mins (until I stopped it). Think I'll stick with two.

I'll take your advice in regards to the memory stuff, thanks :D
 
Bilko123 said:
Woo, excellent stuff. I started a CPU test just to see how it ran on my default settings and on 4 instances of the program (one core assigned to each) it froze up. But on 2 instances of the program it went fine for 20 mins (until I stopped it). Think I'll stick with two.

I'll take your advice in regards to the memory stuff, thanks :D
Click to expand...

I do remember reading on one of the i7 threads that one of the test programs would crash if exactly 4 instance were running (but 8 was ok!). Not sure if thats the same test program, but it could be :)

I normally use Prime95 for my basic testing, and that runs with 4 copies no problem for hours.
 
Oh, btw, you dont mention what cooler you have on the cpu, or what case cooling you have. Keep an eye on your temperatures when the CPU is under load. Temp's are not a major concern in general, but excessive temps can affect stability when overclocking.

My Q6700 peaks at around 75C, but thats still a fair distance from the throttle point on the 65nm Q6xxx series so nothing to worry about, and is 100% stable.

I generally recommend keeping a 65nm cpu at no more than 80C or so, to give a little bit of spare incase of a hot sunny summers day.
 
think the highest rating in win7 is 7.7, but as mentioned, it doesn't make a good benchmark for your system. as you can see. i run a 4890 sapphire vapour x 2gb overclocked @960/1160, which should be no where near your card. but.....
screenshot.png
 
My Seagate 7200.10 scores 5.9 too, and it scores the same either alone, or pair with an identical drive in raid0. Not sure what the disk test is measureing really.
 
Corasik said:
Oh, btw, you dont mention what cooler you have on the cpu, or what case cooling you have. Keep an eye on your temperatures when the CPU is under load. Temp's are not a major concern in general, but excessive temps can affect stability when overclocking.

My Q6700 peaks at around 75C, but thats still a fair distance from the throttle point on the 65nm Q6xxx series so nothing to worry about, and is 100% stable.

I generally recommend keeping a 65nm cpu at no more than 80C or so, to give a little bit of spare incase of a hot sunny summers day.
Click to expand...

Excellent advice, i'm using the stock cooler that came with the chip, although I have been considering getting a better one since I started overclocking. Any recommendations? Saw an Thermalright Ultima 90 which seemed like a decent choice for my setup.

I'm up to 2.62ghz now, not doing badly at all. 3D Mark for the CPU has increased over 300 points already. Lovely stuff. Just need to keep and eye on the temps.
 
Last edited:
You must log in or register to reply here.
Back
Top Bottom