Unigine Valley Benchmark Scores Thread + SUPERPOSITION ***Over 1000 Submissions!***

[Simon771]

Just received my new R9 390 nitro today.

 

My specs: i7 4790 stock (3,8GHz when all 4 cores are active)

R9 390 nitro with backplate: OC to 1150MHz core, 1600MHz VRAM, +50mV, +50% power limit

Z97 Anniversary

Samsung EVO 850 250GB

16GB HyperX DDR3 1866MHz CL10

 

Score: 69,4fps / 2904  Extreme HD

 

[Carde]

On 19-2-2016 at 9:10 PM, Jumper118 said:

because its fun and its free performance. why not have an extra 10% fps? it doesn''t cost you anything. 

Well it costs you some extra power

 

But really getting a 5-10% OC is really a why not get it scenario, going beyond that is just seeing how far you can push it, but it might kill cards. Turns out both my 7970 Lightning BE cards are damaged by me pushing them to the edge, they get unstable after about 2-3 hours workload.

 

That said I still wish the Fury cards had more headroom.

[Jumper118]

2 hours ago, Carde said:

Well it costs you some extra power

 

But really getting a 5-10% OC is really a why not get it scenario, going beyond that is just seeing how far you can push it, but it might kill cards. Turns out both my 7970 Lightning BE cards are damaged by me pushing them to the edge, they get unstable after about 2-3 hours workload.

 

That said I still wish the Fury cards had more headroom.

Well that just a 7970 be thing.  They are all shit.  I've had 3 and they might as well have been bricks. 

[Carde]

3 hours ago, Jumper118 said:

Well that just a 7970 be thing.  They are all shit.  I've had 3 and they might as well have been bricks. 

 

Well any card would give in at some point and I really pushed these 2 to the ragged edge. That said for some reasom the memory seems to be a trouble point with these cards, including these two MSI cards I had 1 powercolor card fail on me due a memory issue got that one replaced under warranty and sold the replacement, that card gave me a bad taste for powercolor.

 

Had a sapphire unit that one never had any issues.

 

Beside those I had to return 1 R9 290 and a 780 for fan failures (bearings on both) and a 780 for just saying fuck this life and deing on me

 

Not bad for 4 years I say

 

Best OCing card I ever had was my Titan though with a modded bios, thing ran 1200mhz stable something like 77-78fps in valley with those drivers, regret selling that card for sure.

[Jumper118]

2 minutes ago, Carde said:

 

Well any card would give in at some point and I really pushed these 2 to the ragged edge. That said for some reasom the memory seems to be a trouble point with these cards, including these two MSI cards I had 1 powercolor card fail on me due a memory issue got that one replaced under warranty and sold the replacement, that card gave me a bad taste for powercolor.

 

Had a sapphire unit that one never had any issues.

 

Beside those I had to return 1 R9 290 and a 780 for fan failures (bearings on both) and a 780 for just saying fuck this life and deing on me

 

Not bad for 4 years I say

 

Best OCing card I ever had was my Titan though with a modded bios, thing ran 1200mhz stable something like 77-78fps in valley with those drivers, regret selling that card for sure.

the be's are known to be crap all mine failed in a couple of weeks and one was doa

 

titan was a monster and i still want one. 

 

[arnavvr]

On 2/23/2016 at 1:11 PM, Jumper118 said:

the be's are known to be crap all mine failed in a couple of weeks and one was doa

 

titan was a monster and i still want one. 

 

GL finding one. Then go for upwards of 450 on EBay. They really should be priced just under the GTX 780Ti, around 215 or so. The Titan was essentially a GTX 680Ti

[TheWoollyMammoth]

CPU: i7 870 at 4,045 GHz

GPU: Gigabyte G1 Gaming 980Ti Core at 1339GHz, mem at 1775MHz

Extreme HD score 3784

 

edit: I'll try again somewhere in the near future.

 

edit2: Okay, figured it out.

[ybriK]

7 hours ago, TheWoollyMammoth said:

CPU: i7 870 at 4,045 GHz

GPU: Gigabyte G1 Gaming 980Ti Core at 1339GHz, mem at 1775MHz

Extreme HD score 3784

 

edit: I'll try again somewhere in the near future.

 

edit2: Okay, figured it out.

I wonder how much performance gain you'll get with a Haswell Refresh/Skylake build knowing that Valley Benchmark isn't really CPU bound.

[TheWoollyMammoth]

9 minutes ago, ybriK said:

I wonder how much performance gain you'll get with a Haswell Refresh/Skylake build knowing that Valley Benchmark isn't really CPU bound.

I think I'd propably get somewhere around 500-700 points more. Max. But that's just guessing. But, maybe we'll find out come autumn.

[Darkman]

33 minutes ago, ybriK said:

I wonder how much performance gain you'll get with a Haswell Refresh/Skylake build knowing that Valley Benchmark isn't really CPU bound.

Valley is CPU bound, it only uses a single core. On my 8120 my 680 would get a score of like 18xx, on my Xeon it got 2073 with the same settings.

[ybriK]

16 minutes ago, Darkman said:

Valley is CPU bound, it only uses a single core. On my 8120 my 680 would get a score of like 18xx, on my Xeon it got 2073 with the same settings.

No, because modern intel chips have better IPC compared to modern AMD chips, that's why you have a higher score when you compared your Xeon against the 8120. Intel = strong single core performance. 2700K vs 3470 very similar single core performance.

 

http://imgur.com/a/cxbYm

 

http://cpu.userbenchmark.com/Compare/Intel-Core-i7-2700K-vs-Intel-Core-i5-3470/1985vs2771

 

http://www.cpubenchmark.net/compare.php?cmp[]=881&cmp[]=822

[FirstimeFailboy]

PABLO

CPU: 4690k @  4.5Ghz

GPU: EVGA GTX 970 ACX 2.0

          1528 Core

          1928 Memory

AVG: 65.5

Score: 2740

[Colomon84]

Colomon

 

CPU: AMD FX-9370 @ stock 4.4 ghz

GPU: EVGA 980ti SC Reference

Core Clock: 1252 Mhz  Boost: 1340 Mhz

Mem Clock: 1903 Mhz

ASIC: 67.3%

[TheWoollyMammoth]

7 hours ago, Colomon84 said:

Colomon

 

CPU: AMD FX-9370 @ stock 4.4 ghz

GPU: EVGA 980ti SC Reference

Core Clock: 1252 Mhz  Boost: 1340 Mhz

Mem Clock: 1903 Mhz

ASIC: 67.3%

Oh, wow. Did you have any background programs running? I mean, I think you should get higher scores than I did with my setup. Or does that cpu really bottleneck that bad in single-threaded workloads?

[ybriK]

17 minutes ago, TheWoollyMammoth said:

Oh, wow. Did you have any background programs running? I mean, I think you should get higher scores than I did with my setup. Or does that cpu really bottleneck that bad in single-threaded workloads?

The AMD FX series don't really have good single threaded performance, they shine better on multi-threaded loads.

 

http://cpu.userbenchmark.com/Compare/Intel-Core-i7-3770K-vs-AMD-FX-9370/1317vs2005

http://www.cpubenchmark.net/compare.php?cmp[]=2&cmp[]=1989

[Raudi_]

19 minutes ago, TheWoollyMammoth said:

Oh, wow. Did you have any background programs running? I mean, I think you should get higher scores than I did with my setup. Or does that cpu really bottleneck that bad in single-threaded workloads?

AMD has piss poor single threaded performance which is why they perform worse the alot of LGA 775 cpu's 

[Lays]

On 2/27/2016 at 1:50 PM, ybriK said:

I wonder how much performance gain you'll get with a Haswell Refresh/Skylake build knowing that Valley Benchmark isn't really CPU bound.

With a 980 ti, Valley can become CPU bound for sure.

[dfg666]

.......

[SebasvanO]

CPU: i5-4460

GPU: Asus GTX 760

Preset: Basic

Average: 88.1

Score: 3685

 

 

Preset: Extreme HD

Average: 44.6

Score: 1868

 

[TheElt]

2 hours ago, dfg666 said:

yea if you partner the 980 ti with a CPU that has weak IPC or a dual core (x4 860k, i5 750, etc.). Not really with strong multi-core modern CPUs, load is assigned to Core 0 in majority of the scenes.

When you're pushing past 65-70fps at any settings, the CPU bottleneck becomes an issue in Valley since most of the logic is single threaded. That's why @Benji_w is lower on the leaderboards than he should be, despite his 970 clocking far better than mine.

 

If I dropped my i5 to get the IPC of an FX chip, it'd perform about the same performance despite having 4 less threads. Valley is awfully threaded, and it's a shame because it really holds back a lot of cards. Hence why a single 980Ti is only 15fps behind SLI 980Ti's when the gap should be much bigger

[Colomon84]

3 hours ago, TheWoollyMammoth said:

Oh, wow. Did you have any background programs running? I mean, I think you should get higher scores than I did with my setup. Or does that cpu really bottleneck that bad in single-threaded workloads?

 

Literally this is the best score I got out of 4 different runs trying different settings.  I'm using MSI Afterburner to overclock this card.  I really think my CPU bottlenecks this card pretty bad considering all the submissions with i5's and i7's have wayy better scores with the same card. Or my overclocking skills are noob status at best 

 

edit: no, I actually closed every program i had running and also ran my windows 10 theme to high contrast black to avoid some of the beauty of windows 10

[ybriK]

48 minutes ago, TheElt said:

When you're pushing past 65-70fps at any settings, the CPU bottleneck becomes an issue in Valley since most of the logic is single threaded. That's why @Benji_w is lower on the leaderboards than he should be, despite his 970 clocking far better than mine.

 

If I dropped my i5 to get the IPC of an FX chip, it'd perform about the same performance despite having 4 less threads. Valley is awfully threaded, and it's a shame because it really holds back a lot of cards. Hence why a single 980Ti is only 15fps behind SLI 980Ti's when the gap should be much bigger

What happened to the devs? Why aren't they updating the Unigine benchmarking software anymore?

[Jumper118]

1 minute ago, ybriK said:

What happened to the devs? Why aren't they updating the Unigine benchmarking software anymore?

waiting for dx12 

[ybriK]

3 minutes ago, Jumper118 said:

waiting for dx12 

Hmmm.... So if they did update it and managed to properly optimize multi-threading loads does that mean I should see higher scores on my 2700K + 290X setup? Because I am seeing 60-70% CPU Usage on CPU 1 or CPU 0 whatever is preferred.

[Jumper118]

9 minutes ago, ybriK said:

Hmmm.... So if they did update it and managed to properly optimize multi-threading loads does that mean I should see higher scores on my 2700K + 290X setup? Because I am seeing 60-70% CPU Usage on CPU 1 or CPU 0 whatever is preferred.

yes. it needs to be made more intensive as well extreme hd should be 1440p i think.