[SOLVED] Vega 56 Clock Speed

[OllympianGamer]

I dont have a lot of experience with AMDs cards but I've got a vega 56, specifically the MSI Airboost OC 8gb version.
So in GPU-Z and in 3dmark it reports my core clock speed at 1622mhz, that is also the speed it is set to in after burner but I've noticed in games or whatever it only ever reaches about 1575mhz.
Not a big deal, that is supposed to be the boost clock speed anyways as I've seen in reviews but if I set the speed to say 1700mhz with afterburner it still only boosts to 1575mhz, it's like I cant overclock the card for some reason although increasing the memory clock speed does work.
Am I overlooking something obvious? I've increased power by 50% and left voltage on auto, no issues with temps either.

 

[TJ Hooker]

Do you see any difference between default power limit and +50%? What is max temp?

 

[OllympianGamer]

Do you see any difference between default power limit and +50%? What is max temp?

Ok I ran Time Spy a couple of times -
1 -
Power +50%
GPU - 55°C highest temp, Max clock speed seen 1560mhz at 99% Usage.
CPU - 3.8ghz, usage varies between 10 and 25%, max temp on cpu test is 37°C and 27°C in gpu tests.

2 -
Power +0%
GPU - 44°C highest temp, Max clock speed seen 1300mhz at 99% usage.
CPU - Practically the same as above.

MY system is -
CPU - 2600 clocked at 3.8ghz with H80i water cooler
MB - Tomahawk B450
GPU - Vega 56 Airboost
RAM - 16gb ballistix tactical 3000mhz dual channel
PSU - Corsair RM650x

I originally asked because I feel like it isn't running as good as it should, in BF5 for example I get around 80fps on ultra 1080p where as other people pull them numbers at 1440p. In Timespy though I tried it with a 1070 and a 1080 and the vega is just behind the 1080 which seems about right but I still feel like it isn't running to its full potential.
Is wattman better than afterburner for AMD cards?

 

[TJ Hooker]

I like Wattman, but I don't think there should be any performance difference between using one or the other.

It looks like maybe you're already maxing out your power limit even at +50%, which would be why you don't see any benefit from overclocking further. I'd look into undervolting, which may get you some more power limit headroom to work with.

Try running userbenchmark to see if your system performs similarly to others with the same specs, post a link to your results here. https://www.userbenchmark.com/

 

[OllympianGamer]

I like Wattman, but I don't think there should be any performance difference between using one or the other.

It looks like maybe you're already maxing out your power limit even at +50%, which would be why you don't see any benefit from overclocking further. I'd look into undervolting, which may get you some more power limit headroom to work with.

Try running userbenchmark to see if your system performs similarly to others with the same specs, post a link to your results here. https://www.userbenchmark.com/

UserBenchmarks:
CPU: AMD Ryzen 5 2600 - 78%
SSD: Gigabyte NVMe PCIe M.2 256GB - 117.8%
SSD: Kingston SSDNow V300 240GB - 76%
RAM: Crucial BLT8G4D30AETA.K16FE 2x8GB - 105.9%
MBD: MSI B450 Tomahawk

For some reason the gpu test failed, tried it multiple times. I then deleted afterburner and ran it again and it did the gpu test. Might try wattman out, see if I have similar issues.

UserBenchmarks: Game 83%, Desk 69%, Work 62%
CPU: AMD Ryzen 5 2600 - 76.8%
GPU: AMD RX Vega-56 - 91.8%
SSD: Gigabyte NVMe PCIe M.2 256GB - 113.8%
SSD: Kingston SSDNow V300 240GB - 73.3%
RAM: Crucial BLT8G4D30AETA.K16FE 2x8GB - 107.3%
MBD: MSI B450 Tomahawk

 

[OllympianGamer]

I like Wattman, but I don't think there should be any performance difference between using one or the other.

It looks like maybe you're already maxing out your power limit even at +50%, which would be why you don't see any benefit from overclocking further. I'd look into undervolting, which may get you some more power limit headroom to work with.

Try running userbenchmark to see if your system performs similarly to others with the same specs, post a link to your results here. https://www.userbenchmark.com/

Fixed it.
GPU: AMD RX Vega-56 - 94.6%
I didn't realise Wattman automatically installs with the crimson software during driver updates, so basically it was conflicting with afterburner and causing weird performance issues, set that to +50% and it boosts over 1600mhz and doesn't crash the test. Cheers I wouldn't have figured it out if I hadn't ran the userbenchmark software.

 

[TJ Hooker]

Ah, I didn't think there'd be any conflicts unless you were actively trying to use both, good to know!