Awesome, will do once I get out of this hell hole.
Yes.
Sorry, must have missed that :/
Should have the update sometime later this evening.
Maybe people should include their CPU (not sure if it matters a lot though)
i think i'll do
TCFX 290x
(down-clocked, 2 other cards need to be downed as i've already noticed throthling in CF from 1080MHz and stable for 2nd card and 3rd should be 1040MHz)
CFX 290x OC
(once written was with 2nd gpu throttling - stable clock is at 1040; this one will include 1150MHz or 1200MHz if it manages to take it)
290x OC (normal 1150 or 1200 core, not 1400MHz or higher as i have a feeling its the last time i use the card when i do it)
290x stock (1080MHz)
ps. Once posted as titan ye puss was done with 1600MHz +2volt on core and 1800MHz mem +1.2v; and catalyst hacks like overriding AA settings to enhance and enabling Xsupersampling (and having AA being rendered on 2nd gpu) I've attached wrong gpu-z screenshot.
It's in the heaven screenshots derp
Well, time for me to enter...
Sapphire Toxic 270X 2GB
Catalyst 14.12 Omega drivers
Win7 Ultimate 64bit
Oh, it matters alright... The score above is on freshly restarted PC with nothing but Heaven running...
Now, while Twitch and Skype are on, the result tanks to this:
Everything is stock...
I'd throw my 8800 gtx into this, but it only supports dx10.
The test is tailored to support DX10 and DX11. Anything from the 8400 GS to the Titan X can be tested under these parameters.
Please, you think I wouldn't plan for that?
Sure there isn't a IQ difference between the APIs?
The score is based on average FPS. If your card can run it, you get a score. Almost all DX11 features are off anyway, Unigine just grouped DX10 into the DX11 setting.
The GTX 295 nearly meets the GTX 950, a card 6 years, 3 process shrinks, 2 memory interfaces and 1 PCI-E revision newer. I'd say the older cards are scoring fairly.
You don't have to run your 8800 GTX if you don't want to. I've got one somewhere in storage that I'll add eventually. :)
I don't think its as much as older cards scoring fairly, I think its more about newer cards scoring badly.
Tested this with my unlocked fury, it ties with 980ti cards in firestrike yet here it struggles against a 390x.
Really not optimal for 2012 and later type cards imo.
Nobody with 980s or 390s have had problems. I dunno, you could be an annomaly.
If your card scores bad, you don't have to share. You are most welcome to start a Firestrike thread where your card scores better.
125fps isn't exactly bad, just less than I expected, I do notice stuttering here and there and low minimums in the single digit range so I'm pretty sure its a anomaly/driver issue.
AHEM.
That is NOT stable. Out of like 7 runs it's the only one to complete. BUT I WILL TAKE MY CROWN NOW.
1 Like
Okay, so after a good bit of fiddling, I got a stable OC.
Score was MUCH higher on the unstable runs... But I figured I'd be a good sport, and see what my card's real-world overclocked performance was. Finally got it down to this, completely stable, and no artifacting. A good performance bump over my initial, stock clock runs.
Apparently, I got a pretty poor overclocker... :/ But for what it's worth, I'm more than happy with my new machine! :)
(Also, this is with my CPU at stock, and XMP at it's lowest value... I may go fiddle with that a bit tomorrow, see if I get any noticeable gains.)
Pretty underwhelming.
Also the reason Strix cards don't overclock as well as the reference PCB is because they use a significantly lower stock voltage, strix PCB has hotpot easily available though.
May i give a small proposition about the spreadsheet :
You may want to add another column and separate the brand and model...
For example, instead of Sapphire Radeon R9 270X Toxic OC, you may want to have one field, saying AMD R9 270X and another one saying Sapphire Toxic OC... It is getting quite cluttered... And then you can make out 2 things: where in general certain GPU is, compared to others, and what brand and model gives better performance for the same GPU...
... I can push 1500 (also not stable above 65c) from a cold boot in an open case, I'll do it when I get home if I have time. It's surprising how close the scores are between your configuration (6990 and 6970?) And a 980ti.
Quadfire 6970's overclocked. So it's the 6990 running a 972MHz (thank god for the engineering sample having awesome chips) and then two 6970s running at 950MHz.
With all cards at 1.2v my PSU tripped. Kil-a-watt read 988W.