Install Steam
login
|
language
简体中文 (Simplified Chinese)
繁體中文 (Traditional Chinese)
日本語 (Japanese)
한국어 (Korean)
ไทย (Thai)
Български (Bulgarian)
Čeština (Czech)
Dansk (Danish)
Deutsch (German)
Español - España (Spanish - Spain)
Español - Latinoamérica (Spanish - Latin America)
Ελληνικά (Greek)
Français (French)
Italiano (Italian)
Bahasa Indonesia (Indonesian)
Magyar (Hungarian)
Nederlands (Dutch)
Norsk (Norwegian)
Polski (Polish)
Português (Portuguese - Portugal)
Português - Brasil (Portuguese - Brazil)
Română (Romanian)
Русский (Russian)
Suomi (Finnish)
Svenska (Swedish)
Türkçe (Turkish)
Tiếng Việt (Vietnamese)
Українська (Ukrainian)
Report a translation problem
As for overclocking:
1. If you don't have it already, install MSI Afterburner and run it.
2. Check "Unlock voltage control" and "Force constant voltage" in MSI AB's settings.
3. Raise the power limit to the maximum, as well as the temperature limit to the maximum.
Adjusting the fan curve to 100% is recommended.
4. Install and run Heaven benchmark in windowed mode so you can control MSI Afterburner and still watch for graphical glitches.
5. Raise memory clock rate by 5 to 25 MHz at a time, clicking Apply after each change, until you start seeing graphical glitches like random patches of color or odd star shaped blotches where they shouldn't be. When you see something like that, tone back the clock rate until the glitches stop.
6. Reset your memory clock back to default (don't forget the highest value you found) and repeat Step 5 for the core clock.
7. Now set core and memory clocks to the highest value you set and run a stress test in Heaven. If it works without glitches or crashing, good. If something goes wrong, turn back the clocks by 5 MHz at a time until it does work properly without glitches.
That's generally how it's done without actually touching the GPU voltage, and I wouldn't recommend touching voltage at all, especially if one has an NVIDIA RTX card as NVIDIA claims that pushing voltage too much, even if it's slight, can significantly destroy the lifespan of the card. AMD cards tend to run pretty hot as is as well, so I would just leave the voltage alone altogether. When I had a Sapphire NITRO+ RX 580 SE Lite, the max temperature was in the 60s at stock, while even when overclocked my Gigabyte RTX 2080 Gaming OC rarely goes above the high 50s. The 2080 is over twice as powerful, yet performs cooler on the same fan curve because AMD and their board partners tend to cheap out on cooling because they're cheap cards.
1. That is a BIOS switch. It's for switching between BIOSs. It's there for people who want to run custom firmware on cards, it gives them a fallback to another BIOS in case their custom firmware breaks the card. By default it will have two stock "gaming" BIOSs on it.
2. Your card most likely doesn't have a "mining BIOS". If it does, a mining BIOS is focused on efficiency not performance.
3. These cards aren't great overclockers, AMD is already pushing these cards to their max. I would rather undervolt a RX card. You can safely play with the RAM and GPU clockspeeds, just be careful with the voltages, it might be smart to not increase the voltage at all and just see how far you can push the card at stock voltages.
I wouldn't run a mining BIOS, if it has one, for gaming.
Overclocking is pretty simple, but I would see how it performs before trying it. There is nothing you should do before trying it, other than checking to make sure your temperatures are okay in your selected computer case.
Unfortunately I am not using and MSI mobo. Its an Asus Maximus Hero VII. Can the afterburner program work with this?
Overclocking the 580 is not really worth it for the heat generated to get a few extra fps.
The RTX1660ti/GTX1070/GTX980ti would've been a more sizable upgrade.
https://www.techspot.com/review/1822-geforce-gtx-970-vs-radeon-r9-290/
Context ^.
Okay great! Thanks for the help. I will give this a shot.
Don't forget to monitor temperatures at stock first and compare the difference.
Honestly, upgrading to a new card was never part of my plan. I was going to stick with the 970 as long as it was relevant to gaming and then slowly phase into my PS4pro while I keep the PC around for work and multimedia related hobbies. But the crash was so severe that slowly my display drivers were failing me for just watching a single youtube video. I don't have much interest in 4k, making sure I can play my games at 1080p at a respectable 60fps was good enough so this card had a great economic value for me.
Got it. Thanks again!
Modern AMD drivers have decent OCing tools already (watt-man) and afterburner conflicts bad with it.
Even if you *do* get afterburner working its generally inferior for AMD cards as it lacks access to power states, which many AMD cards use.
I know for sure that afterburner is a no go on Vega anf am prety sire its not a great option on RX cards either, but I supose I could be wrong...
Undervolting is the way to go with these cards. I recommend dropping the maximum voltage down to 1100 and see what kind of overclock you can get. After finding a stable frequency, then you increase memory speed. HWINFO64 will show gpu memory errors.
Google "XFX RX 580 max overclock"
or "best settings for XFX RX 580" That's what I've always done with my gpu's.
No, you shouldn't do that, because one person's maximum will not be yours. The silicon lottery determines how well your card will overclock.
All you actually need to do is raise temp limit and power limit, and then raise the clocks 5~25 MHz at a time in MSI Afterburner while running a benchmark like Heaven (clicking the apply every single time you change the clock), until you see artifacts on screen, and then finally dialing it back a step or two for stability.