r9 290 pc scoring too low in firestrike and 'low' fps ?

[WRanG]

Hi guys, i need help,
I build a new pc just before christmas as the one i had was starting to struggle.
But somehow i cant wrap my head around my scores and performance in firestrike and various games, I think they should be higher, i may be wrong tho.

In firestrike on the 'standard' run i get just over 8100 points...
In the same 3d mark run i get 110000 in Ice Storm and 19800 in cloud gate.

In metro 2033 benchmark i get avg 71 fps with all effects on max, in 1920x1080 without adv physx.

Dayz standalone (Pre alpha) i can barely run... it jumps between 10 and 50 fps (if i look into the sky 100+) no matter the gfx settings ?
I know its a pre alpha, but seems fishy

My rig consists of:

Gigabyte ga-970a-ud3
Amd fx 8320 overclocked to 4.5 ghz
16 gb ddr3 1600 mhz (4 sticks of 4gb) (only running 1333mhz?, Maybe 2 of the sticks are 1333 and the other 2 are 1600)
xfx r9 290 - stock speeds and cooler
kingston ssdnow 128gb ssd
2 x 500 sata drives for storage and games,
Seasonic m12II 750w 80+ copper psu
cooling on the cpu is a antec h20 620 i have modded to a open water loop.
and i run windows 8.1 64 bit

Temps on the pc are
Gpu: idle 40 Load 86 like
Cpu: idle 14-25 load ~55

If any of you bright minds have any ideas, or knows somethign i dont, please chime in...

And btw. Hi from Denmark, im Tom

 

[Keysplayr]

What do you think your scores should have been? Where is your point of reference?

 

[WRanG]

What do you think your scores should have been? Where is your point of reference?

When looking on the net. Every other r9 290 scores ~9000 in firestrike.
A guy i chat with on a danish site is scoring much higher than me in all the 3d mark tests. He is running a stock 4770k, i know that processor is better than mine, but what makes me wonder is that his graphics scores in 3d mark is much higher than mine too....

Here is his run : http://www.3dmark.com/3dm/1609570?
and mine http://www.3dmark.com/3dm/2037907

 

[Terminal_Boy]

The DayZ standalone (like the ArmA 2 Mod it's evolved from) is very CPU intensive and not at all optimised for multiple cores.

Intel i5 users seem to fare better than i7 and AMD users.

Turning down the graphics settings can actually add more load to the CPU.

Have a read of this:
http://forums.dayzgame.com/index.php?/topic/154395-can-i-run-dayz-retail-post-here/

 

[BallaTheFeared]

Here is his run : http://www.3dmark.com/3dm/1609570?
and mine http://www.3dmark.com/3dm/2037907

Your graphics scores are virtually identical in Fire Strike.

 

[ShintaiDK]

Try run GPU-Z while you run the test, and see how much it throttle.

 

[Z15CAM]

Hi Tom ;o)

Your Unigine Valley & Heaven scores appear to be good for a Reference R9 290.

Suggest you install the latest AMD Beta Driver, AMD Catalyst™ 13.11 Beta 9.2, and download TRIXX v4.8.2. which allows you up to a 200mV off-set. Enable CCC- Overdrive and Set Power Limit Setting to MAX.

Open TRIXX and set GPU Clock to 1150, memory clock to 1500 and VDDC Offset to 156mV and run Unigine Valley also create a Fan Profile or peg it at 95% - It's noisy but does a fair job. This is a descent starting point for OC'g the R9 290/X Cards.

NOTE: Download GPU-Z .075 and enable the Logging during your Bench run. This will record Frequencies and Temps of GPU and VRM's plus show how the Voltage offset behaves.

Also download MemoryInfo 1005. This will tell you what vRam you running - Hynix or Elpida - Don't expect any more then 1500Mhz from Elpida.

The AMD Reference cooler although noisy does a fair job at cooling the VRM's but not that efficient for cooling the GPU. You can also install MSI Afterburner 300 b17 - This will give you a Print Screen Graph of your Bench Run - AB is good but is limited with a +100mV VDDC1 offset.

Your best Benches will occur when you weigh how the card behaves regarding Temp Vs Voltage with frequency assigned before it throttles and limits your OC.

I've Valley Benched 2998 with my Reference Gigabyte R9 290X with Alpida vRam at +156mV 1200/1500. I'm presently waiting a 280x60x140 rad before I go under water.

Merry Xmas & Happy New Year: Ron

 

[WRanG]

So my scores seems to be fine ?
I dont notice any gfx throttle while in game or in 3d mark. they keep stable at 947 and 1250.

The reason The others are getting more points in firestrike than me are due to there cpu ? or ?

 

[BallaTheFeared]

Yes their cpu is getting over 40% higher score.

Your gpu scores are basically identical though.

 

[blackened23]

So my scores seems to be fine ?
I dont notice any gfx throttle while in game or in 3d mark. they keep stable at 947 and 1250.

The reason The others are getting more points in firestrike than me are due to there cpu ? or ?

Correct. CPU performance has a big impact on these benchmarks. That applies to nearly all of them: Firestrike, Valley, and 3dmark11. It is particularly skewed with 3dmark11: The final score is weighed heavily by physics (calculated by the CPU) and Intel hexa core CPUs basically dominate this benchmark. Although you can simply look at the graphics score alone which isn't affected by CPU as much (with 3dmark11, anyway). FX CPUs don't do nearly as well as the other intel offerings in these benchmarks, but as others have stated - your graphics score is fine. Looks to be in line with what it should be.

 

[Z15CAM]

Regarding CPU: I GFX Bench with an i7 2700k idling at 1600mhz with Core Parking disabled.

 

[lyssword]

yep, intel cpu is better for gaming type of benchmarks, especially at lower detail/resolution settings, even if they are measuring GPU mostly. His i7 4770 has a pretty nice 4.5 overclock as well.