GTX 970 overclock results!

[jty0yt]

So on my GTX 970 I have upped the core my 120 MHZ observing a boost clock of 1479mhz. MY fans are around 50 percent at 100 percent usage and the temps are 52 degrees! I have only upped the voltage by 15 MV and the power limit to 105%. I love this card, give me your results below!

 

[xzerqiin]

xgerziin, using your custom settings I beat your score which is odd as my card is at 1430mhz during the test and your gpuz reports over 1500mhz, be honest did that clock hold during the test?

says on your score report, you used 1680x1050p. that's a lower resolution than 1920x1080p. And yes my clock is still holding up to this time, ran Furmark to it already, total of 7 valley rotation when i tapper the OC down, it hit +250 Mhz and was able to run valley but it can't play the game without putting power limit to 110% which increases my card's volts to 1.25, i want to keep it at 1.2060.

 

[xzerqiin]

Set a standard preset for this Thread:

Heaven:
Mode: 1920x1080p 8xAA
Preset: Custom
Quality: Ultra
Tessellation: Extreme

Valley:
Mode: 1920x1080p
Preset: Custom
Quality: High

Check your max volts with GPU-z too.

Would like to add, lets have a hall of fame score records! 😀

 

[xzerqiin]

Anyone know why i am getting pretty slow load rates?
I mean witcher 3 at ultra loads in 5 seconds. but games like league of legends, batman take a while
I have SLI MSI 970 GTX
Asus Rampave V extreme
32gb DDR
5930k Haswelle intel CPU
At high setting 55 Celsius is my max.

Either you have no SSD
Very weak Powersupply
No proper cooling
Non-realistic Computer
Or wants to show off.

 

[dsr07mm]

why are people running with custom tests? use one of the presets preferable extreme hd. Ok just noticed mine was forced to custom, ok fair enough.

here is mine using my modded v5 bios on a evga 970 FTW.

Because Tessalation is not on Extreme and resolution need to be 1080p for benchmarks if you want to compare them with others.

Thats why people before these posts reported fps of 63-67.
Stock MSI 970 is on 53fps usually, clock around 1471mhz gives you around 57.5fps, clocks around 1520mhz are usually on 60fps. Memory clock could add 1fps per 200-300mhz in this test.

 

[xzerqiin]

Now I've done it.
Placed my CPU OC with GPU OC.

Sadly i needed to drop 10Mhz from my GPU core, because i don't want to increase fan speed. 😛

 

[LastTulare]

I did my benchmarking according to this website thread:

http://www.overclock.net/t/1360884/official-top-30-unigine-valley-benchmark-1-0

Its funny because its actually a top 1000 list not a top 30 like the thread title states.
I followed their rules stated to be considered legit for a spot in the listings, and this is what I got after hours of tweaks and testing. I feel like I should use FRAPs to video capture my benchmarking to prove its validity lol

 

[chrysalis]

sadly unigine lacks the function to render higher than the display (my display is 1050p).

Your card must be one hell of a card if it can stay at 1500+ without throttling during the bench. I had to bios mod my card to get it even to stay above 1400 and using 1.275V.

 

[xzerqiin]

sadly unigine lacks the function to render higher than the display (my display is 1050p).

Your card must be one hell of a card if it can stay at 1500+ without throttling during the bench. I had to bios mod my card to get it even to stay above 1400 and using 1.275V.

Its really a luck of draw. I've had GPUs under 50 deg C on water and cant push higher even with celestial volts. Sometimes, your enemy is not the voltage control but the stability of your GPU's memory and motherboard chipset. Or worst is your PSU.

 

[xzerqiin]

I did my benchmarking according to this website thread:

http://www.overclock.net/t/1360884/official-top-30-unigine-valley-benchmark-1-0

Its funny because its actually a top 1000 list not a top 30 like the thread title states.
I followed their rules stated to be considered legit for a spot in the listings, and this is what I got after hours of tweaks and testing. I feel like I should use FRAPs to video capture my benchmarking to prove its validity lol

Very nice! What clock speeds are you running those babies. 😀

 

[chrysalis]

in my case its the gpu core :/ it wasnt even stable in its stock config. I had to bios mod it to even get it stable at stock.

I will post back with scores if I find a way to get unigine to render higher than my display res. Or use a spare monitor I got which is 1080p.

What I can report here is I am finding bench's to give the same result with 1430 boost as 1480 boost on my card, the reason been the card hits the power limit so throttles back anyway (even with a modded higher power limit, I am now hitting hardware limits of the VRM). 1430 boost needs 1.206 vddc on my card, so I may fall back to my V5 bios which has that VDDC as limit.

 

[xzerqiin]

So i just overclocked my GTX 970 the Gigabyte version G1 edition and the core clock was 2019 mhz and a memory clock of 4305 mhz (first time overclocking and i ran it in valley on HD EXTREME preset and got a score of 1165 seemed stable the temp was around 50 and no artifacts

Pics or it didn't happen.

So i just overclocked my GTX 970 the Gigabyte version G1 edition and the core clock was 2019 mhz and a memory clock of 4305 mhz (first time overclocking and i ran it in valley on HD EXTREME preset and got a score of 1165 seemed stable the temp was around 50 and no artifacts

Pics or it didn't happen.

https://imageshack.com/i/f0JxKGMCj https://imageshack.com/i/ipfvX0ZWj https://imageshack.com/i/f0c1v4Swj

Yeah, I don't think that's an actual, functioning overclock. It looks like your GPU is being throttled to oblivion, which would explain why your benchmark score is less than half what a GTX 970 should be doing.

For reference, these are the numbers I get with the stock-settings on my EVGA FTW+

hmm interesting i didn't know that was a thing although it runs games well enough but yeah i do recall getting a little higher score at lower setting i wasn't sure if my CPU was affecting it at all or not (BTW after doing some quick research doesn't throttling only happen when the gpu starts to get too hot my card is well withing safe temps and if there is no artifacts and everything runs fine i wonder why i'm able to OC it so much even if it isn't functional?)

I think 1 reason why you got a low score is probably (highly possible) an unstable OC and you probably came across a black flash during your Valley bench, valley will still say you are running at 2Ghz but in actuality during that black flash your GPU rebooted to stock settings.

 

[xzerqiin]

in my case its the gpu core :/ it wasnt even stable in its stock config. I had to bios mod it to even get it stable at stock.

I will post back with scores if I find a way to get unigine to render higher than my display res. Or use a spare monitor I got which is 1080p.

What I can report here is I am finding bench's to give the same result with 1430 boost as 1480 boost on my card, the reason been the card hits the power limit so throttles back anyway (even with a modded higher power limit, I am now hitting hardware limits of the VRM). 1430 boost needs 1.206 vddc on my card, so I may fall back to my V5 bios which has that VDDC as limit.

Can you post here your GPU-z. Check which memory clas you got (elpida, hynix, samsung). I've also heard that, yes, EVGA has almost no room for volt bump. Also your cooler could contribute it it cant keep the memory modules cool (they made funny placement of memory modules on the PCB making the others over heat).

 

[chrysalis]

nothing wrong with my memory. It overclocks fine compared to others.

Read this thread I made for full story.

http://www.tomshardware.co.uk/forum/id-2716022/evga-970-ftw-performance-modded-bios-discussion.html

 

[dsr07mm]

Now I've done it.
Placed my CPU OC with GPU OC.

Sadly i needed to drop 10Mhz from my GPU core, because i don't want to increase fan speed. 😛

Why not ? I have
30% - 40c
60% - 50c
70% - 60c
80% - 70c

I never go above 61c. Fans arent going to die that fast as people think. Instead 20 years it will be 13 years. Do you really have in plan to keep card for more then 6 years anyways ?

 

[xzerqiin]

Why not ? I have
30% - 40c
60% - 50c
70% - 60c
80% - 70c

I never go above 61c. Fans arent going to die that fast as people think. Instead 20 years it will be 13 years. Do you really have in plan to keep card for more then 6 years anyways ?

I want it whisper quiet. XD
Everything is underwater too. So CPU/GPU/Ram shares on 7x120mm Rads. If i have GPU OC'd (1.3/1.4 boost/1.55 on actual reading) alone temps go as high as 52c at stress, with CPU OC (4.7) temps go 59c at stress [prime and fur].

 

[LastTulare]

I made a video of my benchmark results that include my settings and what I have inside my rig. Im not sure why the preset on the score changed to custom. I start the test using the Extreme HD preset. I shall investigate this further.

http://youtu.be/He8wUCIOJX8

 

[Sergeant_Sneaky]

I made a video of my benchmark results that include my settings and what I have inside my rig. Im not sure why the preset on the score changed to custom. I start the test using the Extreme HD preset. I shall investigate this further.

http://youtu.be/He8wUCIOJX8

Putting it in fullscreen seems to make it custom for me.

 

[chrysalis]

Tried to get 1080p using gedosato, ended up with a 1840x1150 test which is actually a bit more pixels than 1920x1080 so here is the result.

 

[kuleto]

I have the MSI GTX 970 LE (lite edition), just using MSI Afterburner, I have core voltage at +5, (I guess needs to go +20 to make any diffference), 110 power limit, +147 core clock for 1425, +480 memory clock for 3985, a simple custom fan profile that keeps temps at 68C highest observed. Seems pretty stable, was just playing crysis 3. 3570K at 4.2ghz. Got a 32630 Catzilla score on 576p (resolution) tests, good for 246 place overall. Happy with it, got it for $300 on sale on Newegg.

Update: Bumped up OC on CPU to 4.3 ghz setting dvid to +0.05, and I really liked the results. The responsiveness seems much quicker playing crysis 3 and whatnot, I had it OCed to 4.2 and it was kind of laggy with mouse input and bumped up the dvid and that helped it, but its much better now. I haven't really OCed it much, was at 4.0 normally and like -0.6 or so dvid. Temps are higher, getting in the 70s some, but seems okay, wouldn't want to go further than that. Bumped up the 970 by +3 clock. CPUID HWmonitor showing its using 1.012 V max, dont know how accurate that is though. See if setting dvid to 0.0 will lower CPU temps a tad. This is at 78 F ambient temp. Cards temps were mid 60s. Titanfall was barely stressing the card, my fans only turning on at times with highest settings and 4x msaa v-sync on.

Great card, huge improvement over my previous R9 270X.

 

[xaephod]

lol, gotta say nice edit. almost believable XD now if only you edited the bigger score to the bottom pic with the higher avg FPS. i mean really? how does 65FPS avg score higher then 82FPS avg? XD nice try trying to troll.

At minimum, his bottom score is legitimate with his numbers and score being very close to mine. But I don't see an Extreme HD preset anywhere.

For AAx8 1080p, Ultra and Extreme Tessellation, I get:

I just ran the benchmark at the same settings with a 2500K oced to 4.3. I scored 2000. How did I best your core? Your cpu is a lot better?

 

[Openupitsdave]

On Valley Benchmark I got my EVGA 970 FTW card to achieve 65.4fps without any artifacts using this OC:

+40mV
1447mhz (boost)
1953mhz memory clock
power limit 110%

Valley settings were maxed for 1080p


Soo.... I see alot of people getting 1500+mhz boost clocks, where I only got 1447mhz... I may have gotten a crappy card..

What do you think?

 

[LastTulare]

Soo.... I see alot of people getting 1500+mhz boost clocks, where I only got 1447mhz... I may have gotten a crappy card..

What do you think?

Whats your ASICS score?

 

[Openupitsdave]

GPU-Z says 70.2%

 

[xzerqiin]

EVGA GTX 970 SC ACX 2.0. 110% Power, +90 +120 (No voltage changes) Can't push any further.

Past that I just crash out of everything. Heaven would crash my entire Pc when pushed past that. Not sure if it's something to do with my build, or I just got a rubbish overclocker.
Any ideas?

Had two previous GPU's in this build, a Gigabyte GTX 970 Windforce about 9 months ago, and more recently a Sapphire R9 290 Tri-X. Both overclocked like garbage. Is it my PSU, Motherboard, or am I just unlucky?
I have a general understanding of overclocking, no real advanced overclocking experience, but something just doesn't seem right here, when comparing my results with everyone else.

Some more info.
110% Power / +90MHz Core / +0MHz Memory = Past this, crashes on Heaven and FireStrike
110% POwer / +0MHz Core / +150MHz Memory = Past this, crashes on Heaven and FireStrike
Actual Numbers: +110% +90 +120 = 1470Mhz / 3623MHz (Read using MSI Afterburner)
ASIC 74.8%
Memory Type: Elpida
Bios Version: 80.04.36.00.70

Try to not move the power and keep it 100%, don't touch the volts either. Open your gpu-z. Do a benchmark with your cards stock clocks. Check how much Volts the card eats. Then try cranking the core up by +10 (amd) or by +40 (nvidia 900 series). Don't touch the memory. Also 1 problem that you got with your card is the memory (Elpida) bad overclocker and can't sustain heat [bad for high core clocks]. Either Hynix or Samsung. Samsung is best.

 

[Openupitsdave]

What does everyone use to read their boost clocks?? GPU-Z, Valley, and Afterburner all give me very different readings.

As you can see, all 3 programs are giving me different numbers for both boost clock and memory clock.. heres the numbers if you cant see it in the pic:

boost/memory
Valley: 1606/3920
GPUZ: 1442/1960
Afterburner: 1493/3920

Which is correct?