TerraRaptor's 21 min 55 sec 344 ms SuperPi - 32M run with Mobile Athlon 64 4000+ (Newark)

[d0minat0r]

Competition is over, we know who is winner.

Edited March 5 by d0minat0r

[Mr.Scott]

Powerhouse!

[Mr.Scott]

What vcore?

[TerraRaptor]

38 minutes ago, Mr.Scott said:

What vcore?

1.78v here. Cpu is happy with 1.6v @3.55ghz and voltage curve goes to sky after that. I also wasn't able to validate this cpu above 3.65ghz this time - no matter voltage. I'm going to check some other things later this month (vdda and ldt voltages).

[TASOS]

What can i say.
After so many years (in this hobby) , i thought i've seen it all.

But ... i was wrong.

Congrats for this awesome combo (board and cpu) and score !!

Die temp ?

 

[TerraRaptor]

Should be -30 with the load like that (knowing my ss unit). -40 when idle. I control temps only with hot chips that may de-frost evaporator in order to shut down just before it happens.

[TASOS]

-30C die temp or evaporator temp ?

 

[TerraRaptor]

Evap. Have no idea of the die temp? Is software monitoring that good with cold? 

[raiden]

Nope, not really. You can measure delta between load and idle pretty accurately but I’ve had temp readings colder than my evap temp on A64.

[unityofsaints]

Epic score! Platform record.

[TASOS]

On 3/6/2025 at 7:19 AM, TerraRaptor said:

Evap. Have no idea of the die temp? Is software monitoring that good with cold? 

I asked about die temps , because you know how weird these E5 and E6 cores react with negative temps.
No software monitoring.
Many years back , when i was using extreme cooling , i remember i had to actively monitor realtime the die temp with an extra temp probe on the Fluke.

We had temp probes with very thin wires , that you could navigate through the cpu pins.
At the edge of stability , even a 2C temp difference had an impact on benchmarks.

For example , at -8C die temp you could finish the run , while at -10C you were blue screening (temp related) cold bug.

When using an SS it's very hard to hold the temps you need.
I was never lucky enough to have a 754 or 939 cpu without a cold bug (and we didnt have a list of cpu steppings to have ... back then).

On 3/6/2025 at 4:31 PM, raiden said:

Nope, not really. You can measure delta between load and idle pretty accurately but I’ve had temp readings colder than my evap temp on A64.

What do you mean ?
Evap temp will always be higher (colder) than die temp

[unityofsaints]

On 3/8/2025 at 12:08 PM, TASOS said:

(and we didnt have a list of cpu steppings to have ... back then)

Is this list of CB-free steppings available somewhere?

[TASOS]

On 3/10/2025 at 2:41 AM, unityofsaints said:

Is this list of CB-free steppings available somewhere?

People that are still benching 754-939 with xtreme cooling should have a clue (i guess)

It wasn't such a big list anyway.

I can't find my notes at the moment (must be stashed away in some board box with other stuff ... probably)
and as always , i will find it when i wont be looking for it.

[ObscureParadox]

Not CB free, but this was -130c ish - https://hwbot.org/benchmarks/pifast/submissions/4569395