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

I have decided to go Dothan

Overclockers is supported by our readers. When you click a link to make a purchase, we may earn a commission. Learn More.
Gautam said:
Use CPUMSR. Hard to find a good download link, that's the only problem.

My mount appeared to be the problem, lost about 10C doing it again. And I still didn't do a great job. I did a 32M superpi at 237x11. Matched A64's about 300MHz faster at 25' 57".
Click to expand...
CPUMSR
lladrv.dll for the CPUMSR program
Found these for someone who was looking for that DLL in the windows OS forum, figured I'd throw them here since you mentioned it was hard to find :)
 
The Battlefield 2 demo has driven my needs for RAM to insane heights...only to imagine what the full verison will need.

I am tempted to get 4x512 BH5, in what way will that limit my OC options?
 
Gautam said:
From what I hear, not too much. That's just droolific though....4x512...
Click to expand...

Well, the Twinmos Speed premium 512 BH5 costs only 58€ at my favorite online store so i am very tempted to get them as soon as i can gather the money.

Plus, i want to see what the upgrade 1GB -> 2GB does for BF2, other games and windows use.
 
Can you explain to me why getting 4x512 rather then 2x1gb is better for performance. I alway's thought the less dimm's the CPU had to interact with, the faster the data thorough ouput.

Correct me if I'm wrong, just trying to learn:)

PEACE
 
Sjaak said:
Well, the Twinmos Speed premium 512 BH5 costs only 58€ at my favorite online store so i am very tempted to get them as soon as i can gather the money.

Plus, i want to see what the upgrade 1GB -> 2GB does for BF2, other games and windows use.
Click to expand...
Be forewarned, UTT does not do well at all in Intel systems. You'll probably wind up a hair above 200MHz. But at this point, the sheer amount of the memory totally outweighs whatever speed deficit it might incur.
 
EWBatOVAclockin said:
Can you explain to me why getting 4x512 rather then 2x1gb is better for performance. I alway's thought the less dimm's the CPU had to interact with, the faster the data thorough ouput.
Click to expand...

With intel systems, the CPU doesn't communicate with the modules, the Northbridge's memory controller does. I am not sure about the exact difference between 2 - 4 sticks and 512 - 1GB, but in general more MB and more sticks = less OC. The BH5 is only available in 512MB sticks, so i have to go that road.



Gautam said:
Be forewarned, UTT does not do well at all in Intel systems. You'll probably wind up a hair above 200MHz. But at this point, the sheer amount of the memory totally outweighs whatever speed deficit it might incur.
Click to expand...

Are those Twinmos sticks UTT? And still, 200 - 225FSB is enough for the Dothan to go 2.4 - 2.7Ghz.
 
I'm about to try the new TwinMOS SP BH5 on my intel rig in my sig that I'm working on. I have to do vdimm, vtt mods, but I want to try to get at least 220fsb 2-2-2... not possible? pfff... I'm sure it is!
 
well with no mods and 2.85v with tcc5 in single channle i can get 2335 with dual i have to use 2336 @220 if i remember correctly... my pics will be up shortly... just a few more things to retest. @2.4ghz with dual channel mem i now score 30secs with 2225@ddr400.
 
Evilsizer said:
well with no mods and 2.85v with tcc5 in single channle i can get 2335 with dual i have to use 2336 @220 if i remember correctly... my pics will be up shortly... just a few more things to retest. @2.4ghz with dual channel mem i now score 30secs with 2225@ddr400.
Click to expand...

I think we have the same ram. This ram performs the best at 2.5-3-3-8. There is actually a very small drop in performance of you decrease the 8 to 6. There is no loss in performance increasing the cas to 2.5 from 2. 2.5-2-2 = 2-2-2 in every test I've done so far. This includes nearly identical scores in super pi and sisoft sandra. I've tested this extensively using multiple chips, including the dothan and the prescott. I have benched and burned this ram at 2.5-3-3-8 using 2.85v at 270fsb with PAT enabled. I think it can go higher, but the dothan is limiting me right now. It is definitely memtest stable at 2.5-3-3-8 @ 260fsb. I couldn't test higher speeds with memtest since my prescott was limiting me. And obviously, you can't use memtest at these speeds with the dothan. The higher fsb tests were done at multi 9 using eist and clockgen.

Try these settings, you may be pleasantly surprised. However, the new TCC5 in the el rev2 plats don't appear to clock quite as well as the older TCCD, since many of the better overclockin' chips were binned out into more expensive OCZ product lines. My chips were one of the last batches with TCCD.
 
Last edited:
Sjaak said:
Are those Twinmos sticks UTT? And still, 200 - 225FSB is enough for the Dothan to go 2.4 - 2.7Ghz.
Click to expand...

I tried the Twinmos SPII UTTs for a while and got some nice results. These hit 2.5-2-2-8 at 250-260fsb dual channel using 3.6v on a ddr booster. This was the best out of six sticks of ram. The range on single channel was between 235-265 at these timings. Uh, that's quite a range- the UTTs are a gamble.

I sold them since the timings didn't increase the performance enough to warrant pumping 3.6v through my board. I'm now running cool and quiet with about a 2% loss in performance using these TCCDs. I normally, will do anything for performance but the reliability of the OCZ booster was making me extremely nervous. I wanted fast 24/7 settings and pumping 3.6v through the board 24/7 is a bad idea.

IN CONCLUSION. I think for the Dothan you should break out the old BH-5 or try some of the new BH-5s. The UTTs, just don't perform as well as the BH-5s; they fly on the AMD boards, but struggle along on intel.
 
Last edited:
well dat i really dont know since i wanted tight timings and i could be happy at 2.4ghz since it drop 2secs in super with 2225@ddr400 went from 32 to 30.

well here is what i got tring single vs dual
To start off with pics!

now with dual channle:
2.4
http://img.photobucket.com/albums/v733/Evilsizer/ocforums/pm24.jpg
2.6
http://img.photobucket.com/albums/v733/Evilsizer/ocforums/pm26.jpg
2.7
http://img.photobucket.com/albums/v733/Evilsizer/ocforums/pm27.jpg

*edit*
i cant see my pics so i changed it to url's. :(
*edit2*
had to make room so no more single channle pics, sorry.
 
Last edited:
Last edited:
you guys think i'd be able to hit 3ghz with phase change?

edit acutally just fridged water.
 
Last edited:
Evilsizer said:
well dat i really dont know since i wanted tight timings and i could be happy at 2.4ghz since it drop 2secs in super with 2225@ddr400 went from 32 to 30.
Click to expand...

I agree with you man. Tighter timings helps tremendously with the Dothan. If you can get tight timings at high fsb, go for it. What I'm saying is if you've got TCCD you may want to try a high fsb with looser timings. At 2.5ghz using 2-2-2-8 timings with 227fsb I got a pi score of 30.3. At 2.5ghz using 2.5-3-3-8 timings with 277fsb I got a pi score of 28.8. This chip sucks so I can't test at any higher than 2.5. It's barely stable at 2.5 :(
 
datura3 said:
I agree with you man. Tighter timings helps tremendously with the Dothan. If you can get tight timings at high fsb, go for it. What I'm saying is if you've got TCCD you may want to try a high fsb with looser timings. At 2.5ghz using 2-2-2-8 timings with 227fsb I got a pi score of 30.3. At 2.5ghz using 2.5-3-3-8 timings with 277fsb I got a pi score of 28.8. This chip sucks so I can't test at any higher than 2.5. It's barely stable at 2.5 :(
Click to expand...
that sucks... i know mine has the tcc5 chips the lable has the v1.1 on it. I really like this new setup... so much quieter then the water cooled rig but the loudest fan is the gpu fan. 2.7ghz(still on asus cooler) is nice but now im using the 6600gt so im going to bench 01se and 03 abit later tonight or tomorrow... i feel like playing some games right now. this is one sweet chip!
 
Drec said:
you guys think i'd be able to hit 3ghz with phase change?

edit acutally just fridged water.
Click to expand...
Well I did on my Vapo LS, maybe 3050 is doable, haven't tried it yet. Can set it with Clockgen at 3100 for screenie, but then froze when I set it to 3200. I got about 27 seconds in pi at 3000 but it was with stock multi (15), so low FSB, and my RAM timings do suck (Rated at 4400 @ 3.0, will do 2.5 @ DDR 200-ish).

About RAM timings: OK, I have realized that they obviously affect pi scores on the Dothan. But what about the 3DMarks? Anybody do any timing comparisons?

Then, as reported earlier, I got the blue screen in Windows then I couldn't boot into Windows above 200 for some reason (maybe it just got too hot in the room and I didn't give it time to cool). Thought I hosed Windows, but it seems OK now.

On Tuesday night, I finally got it booting @ 200+ again OK, and started working the multis, my best being 11. I kind of have a conundrum at 10, as that will be 300 FSB, which the board might be able to do, but the RAM will stop at ~284, so I'll have to drop the multi to asynch. But higher FSB trumps RAM speeds.

Didn't get to try that, as the Vapo blew a fuse or something, dead as Mussolini. :mad:

I did get over 31K in 3DMark 2001, but that was with a 6800GT at 400/1000, basically stock ultra speed. Unfortunately, I have yet to be able to pop in my X800 XTPE with the cascade on it, due to this latest setback. But I think with the ATI on Cascade, I should break 40K on this easily.

BTW, the Dothan is strange in 2001. In some tests, even in certain areas of tests, like car chase, it just smokes (top frame rate was like 721). I actually hit like 770 top frame rate in Lobby low! In others, not so hot. Weird.

And of course, the Dothan will get it's turn on the -85C cascade too, but since the ATI card is already all bundled up with anti-condensation wrap on the cascade, I will run the 2003/2005 benches first (since they are less CPU-dependent).

One interesting thing about the Vapo on the Dothan that should have been obvioous but still surprised me: Even under heavy load, the reported Vapo temps barely drop at all, unlike the Prescott space heaters, which drop the Vapo temps considerably under load, LOL. I'm running 3DMark 2001 and the Vapo says "-50C". :clap:

If you want to see the Vapo mounted, see this thread
 
You must log in or register to reply here.

Similar threads

D
New Build questions
Replies
9
Views
220
Droidriven
D
N
CPU unstable no matter what I do, and it's getting worse
2
Replies
25
Views
1K
mackerel
mackerel
pinky33
Help pick out a laptop please
Replies
6
Views
339
Flamethrower1972
Flamethrower1972
Z
10 to 15 degrees rise in core temps less than a week after delidding CPU
Replies
8
Views
260
MaddMutt
MaddMutt
mackerel
X299 and 7980XE in 2024
2 3 4 5
Replies
93
Views
3K
freeagent
freeagent
Back