O/C opteron 146

[farooqm]

2T because it's an Asus board like mine, won't go above 240Mhz 1T... gord knows why Asus don't fix this one, especially as it's not an issue with most other manufacturers...

You should be able to go higher with 2T if your mem can take it. Otherwise use a divider to keep your mem down while you clock your board and cpu higher.

[jimborae]

i wont jim lol, im using 2x512 geil sticks in dual channel, i hit 2500 stable with 250 mhz and htt x 4 and cpu x10 without touching voltage or ram sticks but when i touched my ram sticks changed it from 2t to 1t timings i then had problem booting up and had to reset the bios using the jumpter on the mobo to actually get it to post. my cpu is on 1.48v atm and im 3 optons away from top voltage in the drop down box... wondering what i should do to be able to get to 2.6 2.8 ghz stable? i mean my pc is fast lol but i like the ideaof it being faster haha nd the knollege behind overclocking is intresting and finally can any1 recomend some programs to check my stability i have prime 95, that benchmarking program mark05 or somign, speedfan and cpu-z...

I think you cant run 1T at high fsb's because of the ram you're using. ie value ram. but as Buff says in real world apps it makes naff all difference.

If you're using air cooling keep a close eye on the cpu temps, you're doing the right thing by being cautious about upping the cpu voltage. Personally I wouldn't put any more that 1.5v - 1.6v through it without really good air cooling and even then I'd hesitate. My overclocks with that cpu were done using high end watercooling.

As for stability programs I use Prime 95, Super Pi, 3dmark 05, Quake 4 & Folding@home.

[iranu]

I'm running Geil Value ram on my bro's pc.

Asus A8R-MVP mobo with 260fsb. Memory is on a 183Mhz divider (91.5% of 200). Memory is running at 236.4hz @ 2.5-3-3-6 1T.

I think your issue is the memory divider. Try lowering it to 133/166.

[Destroyer^]

i really cant find the memory devider :s

[project187]

yer i have been told bout the divider anyuone tell me where it is??

[Max Tractor]

I disagree, it's a big bandwidth hit but in real world apps it isn't a big performance hit imo.

http://techreport.com/etc/2005q4/mem...y/index.x?pg=2


While I agree certain aspects of memory setting, like tighter timmings, you may fail to notice it, I think the difference between 1t and 2t is real world gaugeable.

[Dark Horse]

http://techreport.com/etc/2005q4/mem...y/index.x?pg=2


While I agree certain aspects of memory setting, like tighter timmings, you may fail to notice it, I think the difference between 1t and 2t is real world gaugeable.

Something like ~3% in gaming (at normal resolutions and detail levels) which I doubt is noticeable.

From the article you posted:

Conclusions
Although tighter memory timings and a 1T command rate can certainly improve the performance of the Athlon 64's memory subsystem, that improvement doesn't always translate to better application performance. In fact, with the exception of the Sphinx speech recognition engine, moving to tighter memory timings or a more aggressive command rate generally didn't improve performance by more than a few percentage points, if at all, in our tests. Lower latencies only improved WorldBench's overall score by a single point, and performance gains in games were generally limited to lower resolutions and detail levels.

[Max Tractor]

Something like ~3% in gaming (at normal resolutions and detail levels) which I doubt is noticeable.

From the article you posted:

You forgot to include the 30-40% difference in synthetic BM. What are you running 1T or 2T?

[Destroyer^]

can any1 tell me where the ram devider is on a asus a8n sli deluxe bios? i cant find it anywhere

[farooqm]

Heres a screenshot of the dividers for your board:

Go here first:
http://www.amdzone.com/pics/motherbo.../cpuconfig.jpg

Then:
http://www.amdzone.com/pics/motherbo...uxe/memmhz.jpg

and select one of the divider.

[Dark Horse]

You forgot to include the 30-40% difference in synthetic BM. What are you running 1T or 2T?

Synthetic bencmarks mean very little in the real world though. 30-40% is possibly a bit much but yes 2T does give a bit of a hit to memory bandwith mainly.

[BUFF]

You forgot to include the 30-40% difference in synthetic BM.

usually 2T loses ~15% bandwidth to 1T but as a dual channel 939 has more bandwidth than the CPU needs you don't see that difference in real world apps.

[Max Tractor]

usually 2T loses ~15% bandwidth to 1T but as a dual channel 939 has more bandwidth than the CPU needs you don't see that difference in real world apps.

You more that likely would be able to provide some linkage, as with most things online they can be proved. There is no doubt that in bm, especially sandra, 2t takes a huge hit. I have also read of "no real world difference" When it gets to that stage, it comes down to experience, having had to run my current setup at 2t because of a compatibility issue between my board and ram, when I finally resolved this, and got running 1t, the different command rate was noticable, more responsive. What do you run 2t or 1t?

[BUFF]

You more that likely would be able to provide some linkage, as with most things online they can be proved. There is no doubt that in bm, especially sandra, 2t takes a huge hit.

Well, we can use the link that you provided earlier.
Look at the Sandra bandwidth scores - the 2T results are about 83% of the 1T.

Another nice article on the effect of bandwidth on A64 performance
http://www.madshrimps.be/?action=getarticle&articID=325

I have also read of "no real world difference" When it gets to that stage, it comes down to experience, having had to run my current setup at 2t because of a compatibility issue between my board and ram, when I finally resolved this, and got running 1t, the different command rate was noticable, more responsive. What do you run 2t or 1t?

Again using the link that you kindly provided
"Conclusions
Although tighter memory timings and a 1T command rate can certainly improve the performance of the Athlon 64's memory subsystem, that improvement doesn't always translate to better application performance. In fact, with the exception of the Sphinx speech recognition engine, moving to tighter memory timings or a more aggressive command rate generally didn't improve performance by more than a few percentage points, if at all, in our tests."

& not forgetting that running 2T will often let you run extra bus speed that will recover the loss

If you had a compatability issue with your RAM & board then I presume that in order to resolve it you changed more than just the Cmd rate - perhaps those other changes also brought other benefits?

I run either as the hardware/occasion calls for (I've actually got 4 A64 rigs here) but I'm no longer a benchmarker just for the sake of it.
Same as I don't run RAID 0 anymore - to me more trouble than it's worth.

[iranu]

Did you manage to set the divider to DDR333 as per farooqm's post?

If you select DDR400 then ratio is 200/200 =1 or 1:1 ratio or 100% of your fsb, so as you up your fsb you will up the memory speed by the same ratio.

DDR333 is 166 (DRR400 is 200) so 166/200 =0.83 or a ratio of 0.83:1 Therefore ram will be running at 83% of your fsb.

DDR266 is 133 so 133/200 =0.665 or a ratio of 0.665:1 Therefore ram will be running at 66.5% of your fsb.

e.g. use a divider DDR333 and a fsb = 230mhz means your ram will be running at a speed of

0.83x230 = 190.9Mhz

It's this that allows you to use a "divider" to keep the ram at a sensible speed whilst you overclock the chip by increasing the fsb.

[Destroyer^]

kk i get yah ill try that tonight im on my way to an eam atm lol