3700 San Diego vs. 3200 Venice: Round 2!

[s7e9h3n]

the SD can superPI 1M all the way up to about 2.93 @1.71v, Venice can only manage SuperPI 1M at 2.89 regardless what vcore being use.

And you stopped pushing the vcore on the sd because of heat? What happens when you feed the venice the same vcores as the sd in terms of temps?

[Epsilon]

Maybe a weird question but since the temperature is the only thing that's different:

Did you turn off the thermalthrottling ?

[ben805]

And you stopped pushing the vcore on the sd because of heat?

That is correct.

What happens when you feed the venice the same vcores as the sd in terms of temps?

The Venice ran cooler at the same vcore as SD at @1.71v, but whenever the Venice gets up to around 55'C then BSOD kicked in, this Venice has the same heat tolerance as the Winchester I had. On the other hand the SD NEVER BSOD even at 70'C when I ran the Toast to test the maximum heat load, but at 70'C I'd chicken out and shutdown the system.


Maximum heat load without BSOD:
San Diego: 70'C ~ ??
Venice: <55'C

Maximum overclocked speed on air without BSOD:
San Diego: 2.93Ghz @1.71v
Venice: 2.89Ghz @1.71

[ben805]

Maybe a weird question but since the temperature is the only thing that's different:

Did you turn off the thermalthrottling ?

Always have it off.

[ben805]

Now get this, it's going to get even more complicated....the Venice can do SuperPI 32M at 318 x9 @1.71v, the San Diego can't. But the Venice ran out of steam and called it quit altogether and threw in the white towel at >2.89Ghz, higher vcore doesn't help. While in the mean time....the San Diego kept going like the Energizer Bunny up to about 2.93Ghz @1.71v.

And this, is the reason why it made me believe that my SD should have a higher ceiling.

[Da[]San]

Some great work here ben805 ignore the thread trashing

[s7e9h3n]

Now get this, it's going to get even more complicated....the Venice can do SuperPI 32M at 318 x9 @1.71v, the San Diego can't. But the Venice called it quit altogether and threw in the white towel at >2.89Ghz, higher vcore doesn't help. While in the mean time....the San Diego kept going like the Energizer Bunny up to about 2.93Ghz @1.71v.

Ok, it's obvious what's happening on the 32m runs. Basically, the sd is getting too hot to finish the run, but the venice stays cool enough to do so. TBH with yah, I don't think your San Diego is particularly better than your Venice. All things considered, there isn't much difference between a cpu which can do 2.89 @ 1.71 and one that can do 2.93 @ 1.71. The only reason the SD is still running is because of it's higher tdp rating:

Seems to me these cpu's weren't binned for speed, rather by strength of mem controller and ability to dissipate more current (in other words, the ability to run hotter)

[ben805]

oh sh*t!! max temperature for both of them are only 65'C but I had my SD up to 70'C couple times for like 2~3 min, hope I didn't damaged anything I thought my SD has max temp of 75'C

[s7e9h3n]

oh sh*t!! max temperature for both of them are only 65'C but I had my SD up to 70'C couple times for like 2~3 min, hope I didn't damaged anything I thought my SD has max temp of 75'C

Somebody correct me if I'm wrong, but I think that's max case temp

[PetNorth]

You aren't wrong, it's correct

[terrace215]

yes, but case as in "die case", a.k.a the integrated heat spreader.

[s7e9h3n]

yes, but case as in "die case", a.k.a the integrated heat spreader.

hehehe, I should have known that.....and Ben, I pull my hair out over the temps that your mobo reports to you. Also, more often it's voltage that kills a cpu rather than temp.

[ben805]

OK, then my SD was pretty close to the max then

[PetNorth]

yes, but case as in "die case", a.k.a the integrated heat spreader.

not "die", but max "safe" temp value I think.

[ewitte]

OK, then my SD was pretty close to the max then

Things do tend to slow down once you get to a certain limit where stability is borderline.

Eric

[s7e9h3n]

Superpi tends to care more about latency. If I leave the memory timings at default I'd probably get about 31s at 2.8Ghz as well. With them tweaked its 29s. Thats 2.5-3-3-7-1T (10,12 and mostly 3's). It could easily be 27s or even 26s if I could pull 2-2-2-5-1T.

Eric

Ok, I was curious so I browsed through the Spi thread and this is the listing of 26sec and 27sec Spi1m runs:

18. Mauri Antero - 26.00 seconds - windows - P4 560 @ 5118 [/mbot]
[mbot] Mauri Antero26FX53 @ 3152
19. Sucka - 26.00 seconds - windows - FX55 @ 3185MHz
20. PetervanDamned - 26.00 seconds - windows - amd 4000 @ 3150 mhz
21. cpulloverclock - 26.00 seconds - windows - 3000 .09@3046.9MHz Epow 9NDA3 rev2.1
22. Kunaak - 26.00 seconds - windows - P4E 2.66Ghz @ 4.76Ghz
23. ford - 26.00 seconds - windows - p4 640@4500mhz
24. Cranox - 26.00 seconds - windows - A64 Venice @ 3200 mhz 2.2.2.5
25. enduracell - 26.00 seconds - windows - Win2k3 3500 CAA2C@11x292/13_UTT_Mach1
26. Dani - 26.00 seconds - windows - A64 3500 130nm S939 301x11DFI nF4 UltraD2xKHX3500 BH5
27. LardArse - 27.00 seconds - windows - P43.4EE @ 4744, 339 FSB 3:2
28. hallowen - 27.00 seconds - windows - FX53 @ 3176.4Mhz 235.3Mhz FSB 1:1 Cas 2,3,2,6,11,16
29. The Stilt - 27.00 seconds - windows - 4722MHz
30. GomerPyle - 27.00 seconds - windows - FX53@2998MHz 2225
31. pc ice - 27.00 seconds - windows - FX53@3102MHz 2225
32. Hoschi - 27.00 seconds - windows - FX53@3026MHz 2225
33. Niaboc79 - 27.00 seconds - windows - FX53@273*11 (2 2 2 1 7 12 1T)
34. Shidokanjohn - 27.00 seconds - windows - P4 3.4 EE 2 x 256MB Kingston value ram BH5
35. freecableguy - 27.00 seconds - windows - LGA775 3.4EE ES@ 4.42Ghz (260x17) on DFI LP925XT2 w/OCZ PC24200 DDR2
36. Team Bertha - 27.00 seconds - windows - A64 @ 3103Mhz 310*10 ddr @ 282Mhz

Nobody does 27's @ 2.8, actually, nobody does 27's short of 3G. It just isn't possible with our procs/hw.

[ben805]

27. LardArse - 27.00 seconds - windows - P43.4EE @ 4744, 339 FSB 3:2
28. hallowen - 27.00 seconds - windows - FX53 @ 3176.4Mhz 235.3Mhz FSB 1:1 Cas 2,3,2,6,11,16
29. The Stilt - 27.00 seconds - windows - 4722MHz
30. GomerPyle - 27.00 seconds - windows - FX53@2998MHz 2225
31. pc ice - 27.00 seconds - windows - FX53@3102MHz 2225
32. Hoschi - 27.00 seconds - windows - FX53@3026MHz 2225
33. Niaboc79 - 27.00 seconds - windows - FX53@273*11 (2 2 2 1 7 12 1T)
34. Shidokanjohn - 27.00 seconds - windows - P4 3.4 EE 2 x 256MB Kingston value ram BH5
35. freecableguy - 27.00 seconds - windows - LGA775 3.4EE ES@ 4.42Ghz (260x17) on DFI LP925XT2 w/OCZ PC24200 DDR2
36. Team Bertha - 27.00 seconds - windows - A64 @ 3103Mhz 310*10 ddr @ 282Mhz[/I]

Nobody does 27's @ 2.8, actually, nobody does 27's short of 3G. It just isn't possible with our procs/hw.

does that mean I'm the first one to hit 27's with 2.91 Ghz?!?! not sure if the patched version count though

27's @2.8 might be possible if I can pull off a 28x (1.5-2-2-x) x10 with extremely tight timing and high vmem.

[mad mikee]

Since you're all set up, test this and let us know if it works b4 I take stuff apart after work ( is of course just 'n you )

And while you're @ it, does it help all those others out their w/ venices

[ben805]

Since you're all set up, test this and let us know if it works b4 I take stuff apart after work ( is of course just 'n you )

And while you're @ it, does it help all those others out their w/ venices

I'm still at work but will try that 'tweak' when I get home

By the way, that dang Venice had already went back to my brother.

[Crazykooter]

Ive had a sub 28 sec run on non modded Pi for about week or so now. 27.976 or something LOL. Oh and at 2940 mhz

[burningrave101]

LOL. Got cheap? Tests like this is why I always say buy the lowest rated chip in iteration and turn it into $1000 chip. Pyrite into gold is where it's at.

You kinda over spent gettin the 3200 brudda

Buying a 3700+ San Diego over a 3200+ Venice is kinda over spent but going with a 3200+ Venice over a 3000+ is actually a good idea because of getting the 10x multi and the fact that alot of 3000+'s aren't overclocking that great. The 3200+ seems to hit 2.7Ghz+ more regularly.

[xwoodoo]

Maybe i was lucky
Next week i will have 2 San Diego for tests. Hope i will be lucky too.
Very odd to see almost no difference between San D and Venice in 3DK 2001 coz between 4000+/FX55 Vs Winni was at least 1K in 2001SE

[ben805]

Ive had a sub 28 sec run on non modded Pi for about week or so now. 27.976 or something LOL. Oh and at 2940 mhz

With the non-patched version I still can hit 27's but needed slightly higher clock at 2.93Ghz

Maybe i was lucky
Next week i will have 2 San Diego for tests. Hope i will be lucky too.
Very odd to see almost no difference between San D and Venice in 3DK 2001 coz between 4000+/FX55 Vs Winni was at least 1K in 2001SE

My single stage Phase change is almost done and as soon as I get it up and running I'll do more benches....though without the Venice, my brother sold it to one of the member here and bought himself a 3800+.

[s7e9h3n]

I'm still waiting to see 27's @ 2.8

[ben805]

I'm still waiting to see 27's @ 2.8

hehehe.....was pushing too hard and one of the stick took a dump and wave white flag, when the 2nd pair arrive I'll try again