• Competitor rules

    Please remember that any mention of competitors, hinting at competitors or offering to provide details of competitors will result in an account suspension. The full rules can be found under the 'Terms and Rules' link in the bottom right corner of your screen. Just don't mention competitors in any way, shape or form and you'll be OK.

Perfcap Heaven Benchmark VRel followed by Thrm

Soldato
Joined
20 Jun 2011
Posts
3,675
Location
Livingston
I have a GTX 1080 FTW in a FT03. Airflow is limited and within a minute or two of running Heaven, the card hits 80c and begins throttling back the boost clock. The OSD display settings show the perfcap reason to be voltage (VRel) followed by temperature (Thrm).

Was just wondering if anyone can say for sure (lots of threads and various theories make this confusing) but would it be fair to assume boost 3.0 is monitoring the rate at which the card is climbing in temperature, capping the voltage to keep the card within temp limit, ultimately reaching the temp limit anyway then the Thrm throttling kicks in to further bring the temp within target.

I suppose what I'm really asking is does VRel come into play as a result of thermals or is it only likely to happen when the core can't boost any higher at the current voltage?

I suppose I should try re-pasting and increase the airflow as a test but I thought I would ask here first. Would be good to be 100% sure of what is happening either way.

Cheers in advance
 
Trying to get my head around the above, I repasted the GPU and holy :eek:! EVGA did one horrific job of applying the stock thermal paste. I completely underestimated the ACX 3.0 cooler it would seem.. Side of the case off fans 100% max temp 51c, down from 75c-85c.

Quick manual curve edit and heaven is looping 2100+MHz. Happy days :)
 
In precision or afterburner set voltage to 100%. It means allow 100% not +100% dont worry... Move BIOS switch (tiny black on on edge of card) to slave (allows more temp/power target). You should see less voltage limit 1's but obviously your thermal will remain the issue, potentially it will get even hotter with 100% voltage (+5-8c ish).

My FTW's hit 2,139 max with some memory overclock too... Great cards.
 
Yep cheers Smffy. I’m using the slave bios which I’ve reflashed using the FTW2 slave, 11GHz mem clock as standard :)

Seems 2139 is around the limit of my card too. I had it up to 2170 ish but the screen started flickering, definitely not stable.

Boost 3.0 makes about as much sense as boost 2.0 did - not much if you want to max out the card with manual clocks. Perfcaps are all over the place before power target, and voltage becomes a limiting factor.

I’m tempted to try the Strix T4 bios just to see if the voltage limit is preventing the core from going any higher.
 
There is a hard voltage limit on the actual GPU so nothing to gain on pascal sadly... allowing max on the standard BIOS will hit that limit.
 
I had a quick play with the Asus Strix T4 bios last night. It didn't help any.

Same clocks as the FTW bios. Higher voltage, a bug with the power target and messes up fan control.

Without getting into conspiracy theories, it would appear EVGA have imposed limits on the FTW that don't appear to exist on the Classified model. I suppose they need to justify the existence of both cards, and the difference in price somehow.

Looks like 2140 ish is my limit.
 
I had a quick play with the Asus Strix T4 bios last night. It didn't help any.

Same clocks as the FTW bios. Higher voltage, a bug with the power target and messes up fan control.

Without getting into conspiracy theories, it would appear EVGA have imposed limits on the FTW that don't appear to exist on the Classified model. I suppose they need to justify the existence of both cards, and the difference in price somehow.

Looks like 2140 ish is my limit.

I've tried it before and is more just "looking like" its doing something. All bios are locked to a maximum of 1.093v there is no way to go above this without physically modding the pcb.
 
I would need a multi to check for certain, but the ASUS Strix T4 bios did allow for higher voltage compared to the stock EVGA bios. Going by the OSD readout anyway.

It was leaked by someone a few years back after ASUS released it I believe for a sub zero/Ln2 competition.

http://forum.hwbot.org/showthread.php?t=159025

Should allow upto 1.2v but then maybe the PCB modifications are required to get there in conjunction with the bios flash.

EVGA did release a tool on their forums for Classified owners (no pcb mods required) and it does allow higher than stock voltage as confirmed here:

https://www.techpowerup.com/forums/threads/voltage-tool-evga-1080-classified.227761/

It would appear there is some give depending on the bios and card model. I guess without Pascal bios tweaker it’s hard to find any definitive answer.
 
Yeah but Nvidia limited the actual GPU chip to 1.093v from what I understood BIOS etc irrelevant as there is a hardware limit, that’s the definitive answer for pascal unfortunately.

I’m not sure where all the bios tools came from but I don’t think you will have any joy.
 
If Nvidia did limit the GPU core itself to 1.093v, the EVGA classified model wouldn't be able to push 1.2v. It uses the same GPU core as any other 1080, the difference being the PCB design.

It would appear under instruction by Nvidia, all partners cap core voltage @ 1.093v to standardise specification.

This must be done at bios level. Maxwell was also voltage limited but this could be raised by editing the voltage table within Maxwell bios tweaker.

Since Nvidia have gone out of their way to do this deliberately, it leaves the question open as to why?

Perhaps a 1080 with a decent modified bios could approach 1080ti performance? We'll never know without a way to test.
 
Back
Top Bottom