• 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 Radeon RX 5700 XT Owners Thread.

Man of Honour
OP
Joined
21 May 2012
Posts
31,940
Location
Dalek flagship
How are you guys getting on undervolting your 5700 XTs? I've been able to reduce mine down to 1060mV at the stock max-clock of 2059MHz and 786mV @ 1429MHz without any artifacting. Can't really say if it's drastically reduced temps but I've also already done the GamersNexus washer & paste mod.

Card now gets up to 78℃ "temperature" and 95℃ "Junction Temperature", hovering at about 1700MHz

Oh, and here's it getting a little naked time while I was doing the GN mod:
pwFcP8e.jpg


Welcome and on TROH.:)


The Roll of Honour.


Hk6Mrap.jpg


Date 20/07/2019, Make AMD Reference, Aod, Link
Date 19/07/2019, Make AMD 50th AE, BOFH, Link
Date 18/07/2019, Make AMD Reference, gavinh87, Link
Date 17/07/2019, Make AMD 50th AE, LtMatt, Link
Date 14/07/2019, Make AMD Reference, bake73, Link
Date 12/07/2019, Make AMD 50th AE, jrodga2k5, Link
Date 11/07/2019, Make AMD Reference, LtMatt, Link
Date 11/07/2019, Make AMD Reference, docsonic, Link



To get on the Roll of Honour all you have to do is post a pic with your user name on it of your Radeon RX 5700 XT.
 
Caporegime
Joined
12 Jul 2007
Posts
40,413
Location
United Kingdom
So much moaning in the Reddit thread about 19.7.2 I was going to wait for an update. But I'm struggling with stability in-game with 19.7.1 & undervolting:/
19.7.2 is better than 19.7.1, so i would definitely update. DDU in safe mode and then download and install the full 19.7.2 package. (400+mb)
 

Aod

Aod

Soldato
Joined
7 Oct 2004
Posts
8,662
Location
London
19.7.2 is better than 19.7.1, so i would definitely update. DDU in safe mode and then download and install the full 19.7.2 package. (400+mb)
Agreed - I'm seeing much better reliability/consistency with 19.7.2, haven't done a performance comparison but I'm not really fussed about losing 5-10% at the moment
 
Soldato
Joined
18 Oct 2002
Posts
4,333
Game Testing

5700 XT Underclocked & Undervolted (1429mhz 682mv 75w-85w approx.)

Forza Horizon 4 - Using In Game Benchmark.

1080p Extreme Settings - 8 x AA (looks incredible btw)
85fps 90.4fps average, 76.1 minimum, 111.3 maximum
Stutter count 0
Average latency 35.9

1440p settings as above
66fps 67.7fps average, 56.5 minimum, 831.1 maximum
Stutter count 0
Average latency 62.1

Batman - Arkham Knight - Using In Game Benchmark

1440p - all setting to max - gameworks rain and light on
min. 69 max. 158 average. 110

1080p - all settings to max - as above
min 72. max 235 average. 152

Very excited by these results, especially for 75-85watts !

85watt testing continued:

Metro Exodus - Using Actual Gameplay - Ultra Settings DX12 - No Hairworks or Tess. OCAT Capture. Readings taken from prologue and then start of game on top of building.

1080p - 95fps - 11.8ms
1440p - 84.7fps - 14.3ms (checked a couple of times, perf don't seem to drop that hard).

Plague Tale Innocence - Ultra Settings (Incredible looking game) - OCAT Capture - First Two Missions.

1080p - 73fps average, 13.5ms latency
1440p - 50fps average, 20ms latency.
 
Last edited:
Soldato
Joined
14 Jul 2005
Posts
17,615
Location
Bristol
y4mCmMMmppkdEEgDEEhF1Htc4Ql8uzaXYlJoC0O3rK3RiZkOY3Z8T_W7G9OFsAvPs_NOTQ6wypATh_rZQRm3vagrkuHHydoO-l28TK8kk38IY3w1-u2pDnpCh4T9gXOzhcsXwP-p-cIglacrhlcYxTykqYSaTcGX9te4JOoMfrPTt-d2kSh5NWrmV-mN23H2-KzZKBGw4RAww2sns1kt1z7mA



Go team red, or something. I actually have a solid side on the case but thought I'd get a pic with the side off.

Not actually run any games yet (completely new machine), might well run two power cables into the 5700 XT as others have mentioned. I've not had a decent card in years and there was no mention in the paperwork about using one/two power cables so I went lazy.
 
Man of Honour
OP
Joined
21 May 2012
Posts
31,940
Location
Dalek flagship
y4mCmMMmppkdEEgDEEhF1Htc4Ql8uzaXYlJoC0O3rK3RiZkOY3Z8T_W7G9OFsAvPs_NOTQ6wypATh_rZQRm3vagrkuHHydoO-l28TK8kk38IY3w1-u2pDnpCh4T9gXOzhcsXwP-p-cIglacrhlcYxTykqYSaTcGX9te4JOoMfrPTt-d2kSh5NWrmV-mN23H2-KzZKBGw4RAww2sns1kt1z7mA



Go team red, or something. I actually have a solid side on the case but thought I'd get a pic with the side off.

Not actually run any games yet (completely new machine), might well run two power cables into the 5700 XT as others have mentioned. I've not had a decent card in years and there was no mention in the paperwork about using one/two power cables so I went lazy.


Welcome and on TROH.:)


The Roll of Honour.


Hk6Mrap.jpg


Date 20/07/2019, Make AMD Reference, InvaderGIR, Link
Date 20/07/2019, Make AMD Reference, Aod, Link
Date 19/07/2019, Make AMD 50th AE, BOFH, Link
Date 18/07/2019, Make AMD Reference, gavinh87, Link
Date 17/07/2019, Make AMD 50th AE, LtMatt, Link
Date 14/07/2019, Make AMD Reference, bake73, Link
Date 12/07/2019, Make AMD 50th AE, jrodga2k5, Link
Date 11/07/2019, Make AMD Reference, LtMatt, Link
Date 11/07/2019, Make AMD Reference, docsonic, Link



To get on the Roll of Honour all you have to do is post a pic with your user name on it of your Radeon RX 5700 XT.
 
Soldato
Joined
18 Oct 2002
Posts
4,333
Final Fantasy Online - Shadow Bringers Results.

Looks like crap and runs like greased lightning on a 5700 XT. A great piece of propaganda by Nvidia to use the overall score rather than the fps readings.

5700 XT Underclocked & Undervolted (1429mhz 682mv 75w-85w approx.)

'Maximum' Graphical Settings
1080p - 15,266 - As below - Average 106.8fps
1440p - 11,326 - Result 'Extremely High' - Average 76fps.

5700 XT Overclocked, Overvolted with +90 PL - (2204mhz/875 1263mv 250W approx.)

'High' Graphical Settings
1080p - 16,710 Average 126.95fps minimum 53
1440p - 14,747 Average 101.96fps minimum 29

'Maximum' Graphical Settings
1080p - 16,328 - Average 120.5, Minimum 52
1440p - 13,614 - Average 92.6fps

Final Fantasy XV

Looks significantly nicer than FF online. This benchmark is unfortunately infested with Nvidia Gameworks options and it is probably the main reason why this bench gets leaked before most others as it intentionally gimps the AMD GPU scores by approx 22%

5700 XT Underclocked & Undervolted (1429mhz 682mv 75w-85w approx.)

High Settings - Gameworks ON
1080p - 5,878 - 'fairly high' - avg fps. (OCAT 50-60fps 16ms)
1440p - 4,599 - 'fairly high' - avg fps. (OCAT 36-41fps 27ms)

5700 XT Stock Clock (ish) & Undervolted (1904mhz 985mv 125-150w approx.)

High Settings - Gameworks ON
1080p - 7,110 - 'High' - average fps. (OCAT 74fps 13ms)
1440p - 5,617 - 'fairly high' - avg fps. (OCAT 57fps 17ms).

5700 XT Overclocked, Overvolted with +90 PL - (2204mhz/875 1263mv 250W approx.)

High Settings - Gameworks ON
1080p - 7932 - Average fps. (OCAT 80fps 12.5ms)
1440p - 6184 - Average fps. (OCAT 55-65fps 17ms)

Highest Custom Settings - Gameworks OFF

1080p - 9887 - 'Very High' - Average fps. (OCAT 100fps 10ms)
1440p - 7175 - 'High' - Average fps. (OCAT 75fps 13.3ms).

@LtMatt - Almost there with your requested tests buddy, I just needed more data to show where the sweet spot of my card was.
 
Last edited:
Associate
Joined
26 Aug 2010
Posts
1,852
19.7.2 is better than 19.7.1, so i would definitely update. DDU in safe mode and then download and install the full 19.7.2 package. (400+mb)

Better reults with 19.7.2 in Metro Exodus. I was actually able to have a gaming session with a basic undervolt wattman @ 2060mhz 1100mv nothing else changed. Temps reached 82c.

I have an undervolt that matches my stock timespy & firetsrike graphics scores:

power limit +10%
sDX9nWi.jpg

Results after about 20 mins of Metro without any crashing:

2GRRZSc.jpg

So heading in the right direction!
 
Caporegime
Joined
12 Jul 2007
Posts
40,413
Location
United Kingdom
Better reults with 19.7.2 in Metro Exodus. I was actually able to have a gaming session with a basic undervolt wattman @ 2060mhz 1100mv nothing else changed. Temps reached 82c.

I have an undervolt that matches my stock timespy & firetsrike graphics scores:

power limit +10%
sDX9nWi.jpg

Results after about 20 mins of Metro without any crashing:

2GRRZSc.jpg

So heading in the right direction!

Nice work Doc, good results. :)

Bit confused though you say 2063Mhz but your WattMan shows 1932Mhz and the histogram shows a peak core clock of 1943Mhz.
 
Associate
Joined
26 Aug 2010
Posts
1,852
Ah yeah what I meant is that I tried 2060/1100 last night & that worked okay in-game.

So this morning I tried a different undervolt - 1932/1000 & thats the one in the pics :)
 
Last edited:
Soldato
Joined
18 Oct 2002
Posts
4,333
Great work mate, very interesting results.

I am interested to find out what is the maximum stable game clock you can find at the lowest possible voltage whilst having 100% stability and no artifacts with a +50% power limit.

Answer is - 1904mhz/stock vram @ 955mv.

*updated* Sorry I used wrong reading.
 
Last edited:
Soldato
Joined
18 Oct 2002
Posts
4,333
Gears Of War 5 Tech Test Results

1080p - Ultra Settings - VRAM Usage 4.13-4.44GB

5700 XT
1925/stock vram @ 955mv +50 Power Limit - Average frame rate 105.2 fps, Average minimum frame rate (bottom 5%) - 79fps.
1904/stock vram @ 955mv Stock Power Limit - Average frame rate 98.2 fps, Average minimum frame rate (bottom 5%) - 79.2fps
1429/stock vram @ 682mv (75w-85w) Stock Power Limit - Average frame rate 86.8, Average minimum 71.8 fps.

Note: the power limit making a big difference here.

1440p - Ultra Settings - VRAM Usage 4.40-4.67GB
1904/stock vram @ 955mv Stock Power Limit - Average frame rate 73.8 fps, Average minimum frame rate (bottom 5%) - 60.7fps. 13.2ms
1429/stock vram @ 682mv (75w-85w) Stock Power Limit - Average frame rate 60.3, Average minimum 49.1 fps.

1080ti HOF
1974(stock)/vram 5508 (stock) Stock Power Limit - Average frame rate 89.6fps, Average minimum frame rate (bottom5%) 69.2fps. 12ms.

1440p 4k upsample - Ultra Settings - VRAM Usage 5.25-5.44GB
1904/stock vram @ 955mv Stock Power Limit - Average frame rate 38.1 fps, Average minimum frame rate (bottom 5%) - 32.1fps. 24ms
1429/stock vram @ 682mv (75w-85w) Stock Power Limit - Average frame rate 30.8, Average minimum 27 fps. 33ms

1080ti HOF
1974(stock)/vram 5508 (stock) Stock Power Limit - Average frame rate 50.3fps, Average minimum frame rate (bottom5%) 42.8fps. 22ms.

Conclusion, DX12 and async compute really helps Navi against the 1080ti. To match a 1080ti custom under liquid cooling a 5700XT needs a GPU clock of approx 2200-2250 with a bigger power limit and ram clock. I think it is doable once I get my block. I also need to go through the OCAT report as the frame-times looked a hella lot more smoother on Navi than on the 1080ti.

Note: Afterburner reports much more vram usage than what the game does.
 
Last edited:
Associate
Joined
26 Aug 2010
Posts
1,852
More 19.7.2 testing:

ts = timespy graphics score
1850mhz 965mv +3% ts 8477
1850mhz 955mv +3% ts 8500
1850mhz 949mv +3% ts 8499
1850mhz 944mv +3% ts 8498
1850mhz 939mv +3% ts 8503
1850mhz 934mv +3% ts 8492
1850mhz 929mv +3% ts 8494
1850mhz 924mv +3% ts 8504
1850mhz 919mv +3% ts 8511
1850mhz 914mv +3% crashx2/ ts 8508 but artifact noticed with power at stock/

Just had 15 mins in metro at 1850/919mv no crashes, more testing needed though.
 
Associate
Joined
24 Nov 2013
Posts
138
Location
Sweden
Have had some testing of my own on 5700xt and Firestrike extreme. Tested stock cooler and water. With killawatt meter in wall for card power. 110-120w computer power. 7700k at 5.2ghz on aio. Sappire 5700xt( samsung mem)

voltage scaling stock cooler ( set at 100% at :

core/mem volt watts avg real clock avg Powertarget junktion temp(highest) Fsx Gfx score

1860/910 mhz 913mv 110 w 1810mhz 0% 66 degrees 12110
1920/910 mhz 940mv 125w 1870mhz 0% 66 degrees 12520
1960/910 mhz 970mv 135w 1910mhz 0% 66 degrees 12756
2000/910 mhz 1000mv 145w 1945mhz 0% 72 degrees 12932
2070/910 mhz 1050mv 170w 2015mhz 0% 76 degrees 13220
2110/910 mhz 1080mv 180w 2045mhz 0% 81 degrees 13410
2150/910 mhz 1100mv 205w 2100mhz 0% 87 degrees Fail. card fail in scond test when junktion hit 87 degress.

Watercooling. Soft power table mod. water 24 degrees. 31 degrees idle gpu.

1860/910 mhz 908mv 105w 1800mhz 0% 52 degrees 12122 (210-215w of powerdraw at wall :O)
2151/910 mhz 1091mv 190w 2100mhz 0% 62 degrees 13700
2200/910 mhz 1141mv 210w 2150mhz 0% 65 degrees 13920
2250/910 mhz 1220mv 240w 2200mhz 10% 71 degrees 14050
2270/910 mhz 1250mv 270w 2220mhz 20% 78 degrees Fail. card fail in test 2 when junction hit 78 degrees.


Will test cold scaling next. With some cold water.


Power eff at 1860/910 is funny 110w of power from wall at avg. and 1800 core avg boost(stayed at 1800mhz all the time). Hope AMD can fix mem oc for the card. My guess is that samsung gddr6 can do better then 35mhz oc from stock. 9900k wil likly give some better gfx score for the card. This was just an volt and temp test for my card.
 
Back
Top Bottom