Soldato
- Joined
- 29 Sep 2010
- Posts
- 6,180
I wouldn't have thought a 2600 would bottleneck at 1440p, I have a 1600af (basically a 2600) and my cpu usage doesn't go over about 40 percent when playing gears 5.
-
Competitor rules
Please remember that any mention of competitors, hinting at competitors or offering to provide details of competitors will result in an account suspension. The full rules can be found under the 'Terms and Rules' link in the bottom right corner of your screen. Just don't mention competitors in any way, shape or form and you'll be OK.
The RX Vega 56 Owners Thread
- Thread starter Kaapstad
- Start date
just checked, the GPU is at 99% usage. for most of the time its 80+ fps but in some areas of the maps it drops to 60ish and 58 etc. this is with everything on max at 1440
Associate
- Joined
- 17 Sep 2018
- Posts
- 1,476
I've experienced similar, we have the same CPU and GPU, from what I remember though it happens when there's loads of NPC characters on screen, which I believe is CPU intensive. Because IIRC I experienced the same issue at 1080p. It could even be a fault of the game engine. I could be wrong ofcourse. I just don't expect a Vega to struggle with 2016 Hitman, with it's limited graphics
If you look HUB Hitman 2 Benchmark, the 1% low with a Vega 56 is 79
Hi guys, how do my settings look? I'm going for low temperature as a priority while playing 144hz ultrawide Overwatch, BF4, SW BF2.
https://ibb.co/CmvBZQx
https://ibb.co/CmvBZQx
Just flashed my Sapphire Pulse V56 with a V64 BIOS from XFX. Was a bit dubious but several posts on reddit suggested it worked, so took the plunge and it worked very well.
Currently running at 1700 on the core (This drops to mid 1600's in gaming) and 1100 HBM. At stock speeds, Superposition score at 1080p Extreme was around 3800ish, overclocked I'm hitting 4685 so quite a decent increase. Temps top out at mid 70's and the noise is about the same as the USB fan on my desk so isn't too distracting. Certainly isn't audible through my DT770's.
Was considering picking up a 2nd hand 1080ti but for 1440p this V56 slaps everything I play so will see if I can get a year or 2 more out of it.
Currently running at 1700 on the core (This drops to mid 1600's in gaming) and 1100 HBM. At stock speeds, Superposition score at 1080p Extreme was around 3800ish, overclocked I'm hitting 4685 so quite a decent increase. Temps top out at mid 70's and the noise is about the same as the USB fan on my desk so isn't too distracting. Certainly isn't audible through my DT770's.
Was considering picking up a 2nd hand 1080ti but for 1440p this V56 slaps everything I play so will see if I can get a year or 2 more out of it.
Associate
- Joined
- 28 Nov 2018
- Posts
- 337
- Location
- Monstrocity
Nice! Those are some tidy figures.
Think I'll flash my Vega 56 once the warranty is finished (dual BIOS but the second one is locked and can't be flashed, afaIk). If that fails, I may have to punish my credit card!
What was hot spot hitting? That's generally the issue rather than core or HBM temperature.
Unless you put the power limit to +50, the undervolt won't work as desired. The card's software algorithm interferes otherwise, and you get unstable frame times.
This can be confirmed playing GPU intensive games, and 3DMark's framerate stability tests may also pick up on this. The effect is easy to see in something like Star Wars Jedi: Fallen Order, where the difference in frame stability feels like going from a moving train, to standing still.
It's only that obvious once you realise what's going on, but the difference is like night and day.
What does the power limit do? Will it increase my temperatures? Your right, it's not working as intended!
From a technical stand point, it limits the amount of watts (amps X voltage, per second) the card can potentially use, but the drawback is that some parts of a scene use more power than others. The algorithm tries to maintain all of it to the same limit of 165 watts on a stock Vega 56, and does so by using a management algorithm.
The card can have some pretty high spikes in power usage at stock settings, so the algorithm serves a purpose. Once undervolted, the card can only peak so much, but the relative difference can be much higher at lower average power usage. This creates high variance in frame times as the algorithm tries to balance it all out. At that point, The drawbacks outweigh any potential benefits or even necessity.
For testing your undervolt, something like the free Final fantasy XV benchmark is excellent for determining stability and peak power levels. It uses a form of virtual resolution that allows you to test complex scenes at various resolutions on any screen. I recommend running each of the three default benchmark presets at 1440p and 4k in a single sitting to determine stability, while the 1440p standard test in great for establishing constant power and heat during a normal gaming session. For determining the max power the card can pull while undervolted, try the 4K standard test looped a couple of times.
Something like MSI Afterburner overlay or the driver in-game overlay should provide all the required information on-screen, or an open GPU-Z statistics window on a second display works just as well if you have multiple monitors.
Last edited:
So Core gets to 75c, Hot Spot goes to 110. Never heard of the hot spot temp before to be honest, how would you fix this?
Card is probably throttling at that temperature. More air may help, but replacing thermal paste / pads is more effective. I've not actually found a definitive source on what the hotspot temp is actually referencing. See what the hotspot gets to on the stock BIOS as well. This is one of the concerns with the V64 BIOS mod, it can push the card a bit too hard and one way to see that is hotspot temp.
Soldato
- Joined
- 18 Oct 2002
- Posts
- 9,317
Silly question but does that also apply if you have it set to +30, etc.?
Thanks! Would I have to change the thermal pads or is just putting some thermal grizzly on the core enough?
Seems easy to deshroud too...wondering if sticking two 120mm fans under it might work
Just the core paste, pads should be ok.
Only the +50 power limit setting acts as a simple hard limit as far as I'm aware.
In my experience, the algorithm just adds an additional abstraction layer that is largely detrimental in most practical scenarios with significantly reduced voltages. It would be more beneficial to use the method in my last post to quickly determine your particular hardware's average power usage and heat output, and then proceed to set up a profile that best suits your needs and particular setup. The reference design Vega uses a top of the line VRM and PCB design, and can easily handle peak power.