390/290 crossfire tested

[App4that]

This is going to be a long post, but I hope it helps someone.

My PC is as follows.

4690k @ 4.6GHz

GA-Z97X-UD3H-BK

H105 cooler

Gskill RipJawsX 16g @ 1600MHz

Valley and Heaven run off Ultra II SSD

3D Mark Firestrike run off WD Black HHD

Coarsair RM750 PSU

GPU 1 is a Sapphire Nitro 390 OC with backplate

GPU 2 is a Sapphire Vapor-x 290 with backplate

(GPU numbers based off final position in case, not starting order)

All tests were performed 3 times, the reported number is an average of the three results. The clock speed of the overclocked GPU2 was chosen as it would not reliable overclock the memory to 1500MHz to match GPU1. So 10MHz was added to the core, to offset the 50MHz missing on the memory. Drivers where removed and installed before the first tests of GPU2, then after GPU2 but before GPU1, and finally after GPU1 but before the crossfire tests.

GPU2 (290)

Valley (Maxed)

1000MHz (core) Score:2391

1400MHz (Mem) Avg. FPS:57.1 High.FPS:108.6 Min.FPS:29.1

Avg. Temp:76c

Valley (maxed)

1050MHz (core) Score:2500

1450MHz (Mem) Avg.FPS:63.15 High.FPS:120.3 Min.FPS:31.45

Avg.Temp:76c

Heaven (maxed)

1000MHz (core) Score:1371

1400MHz (Mem) Avg.FPS:54.4 High.FPS:108.6 Min.FPS:28.6

Avg.Temp:76c

Heaven (maxed)

1050MHz (core) Score:1435

1450MHz (Mem) Avg.FPS:56.8 High.FPS:114.5 Min.FPS:31.9

Avg.Temp:76c

3D Mark Firestrike

1000MHz (core) Score:9688

1400MHz (mem (11726) (9198) (4359) <----(graphics,physics,combined)

Avg.Temp 61c

Firestrike

1050MHz (core) Score:10103

1450MHz (Mem) (12337) (9193) (4574)

GPU 1(390)

Valley (maxed)

1040MHz (core) Score:2702

1500MHz (mem) Avg.FPS: 64.6 Max.FPS:122.3 Min.FPS:25.3

Avg.Temp:70c

Valley (maxed)

1100MHz (core) Score:2807

1500MHz (Mem) Avg.FPS:67.1 Max.FPS:127.4 Min.FPS:32.4

Avg.Temp:72c

Heaven (maxed)

1100MHz (core) Score:1505

1500MHz (mem) Avg.FPS:59.8 Max.FPS:121.3 Min.FPS:28.7

Avg.Temp:71c

Firestrike

1040MHz (core) Score:10315

1500MHz (Mem) (12551) (9200) (4787)

Avg.Temp:66c

Firestrike

1100MHz (core) Score:10678

1500MHz (mem) (13112) (9194) (4967)

Avg.Temp:66c

Crossfire (stock speeds, which are the lowest tested for each, both ran at their speeds independently)

Valley (maxed)

Score:4427

Avg.FPS:105.0 Max.FPS:182.3 Min.FPS:38.4

Avg.Temp:77c (top card) 74c (bottom)

Heaven (maxed)

Score:2621

Avg.FPS:104.1 Max.FPS:230.5 Min.FPS:29.85

Avg.Temp:80c (top) 76c (bottom)

Firestrike

Score:15517

(22805) (9076) (6654)

Avg.Temp:77c (top) 75c (bottom)

FUN FACTS

Adding all the scores up and seeing what's the difference.

290 to 390 with similar clocks gives a 4% increase in scores.

390 to Crossfire gives a 36% increase in scores.

290 to Crossfire gives a 38% increase in scores.

Case temperatures went from 41c to 50c when crossfire was enabled.

It looks REALLY cool.

[Arty]

 Good Work

[AresKrieger]

Well it worked out, and may the fallout bring you much enjoyment with this new found power.

[dragoon20005]

the scaling of the Crossfire is really close with only 2% difference

 

but strange is that 290 being the top card and CF with the 390 nets the extra 2%

[App4that]

the scaling of the Crossfire is really close with only 2% difference

 

but strange is that 290 being the top card and CF with the 390 nets the extra 2%

I have it set up so most of the time I only run the 390, I'll only turn on the 290 when needed. Thankfully the 390 runs at normal temperatures when crossfire is disabled. I wanted to drop the 290 down to the last slot but that puts it right up against the PSU.

[winningsince1337]

snip

Was thinking of doing this myself. Your RM750 runs both of them no problem?

[don_svetlio]

Was thinking of doing this myself. Your RM750 runs both of them no problem?

a 390 uses about 300W give or take

a 290 does the same more or less

rest of the system i see doesn't need more than 150W so it's fine

[App4that]

Was thinking of doing this myself. Your RM750 runs both of them no problem?

I was very careful to gradually increase the load to see how the PSU reacted. Was a little worried because I had heard the older models of the PSU were slow to activate their fans under severe load, but it's almost instantaneous. So no issues yet. I might replace the PSU just to be safe if I find myself using the crossfire more.

[FellTheSky]

i think your valley crossfire score is wrong. Not even triple 980ti's oced can score that

[App4that]

i think your valley crossfire score is wrong. Not even triple 980ti's oced can score that

Yes, thank you. Goofed up and put the Firestrike score there. Edit in process