The Radeon RX 5700 XT Owners Thread.

[Voodoo_what]

I will support this tangential conversation by adding that I have a 1200w Seasonic PSU that is about 8 years old. Powers my Unraid server 24/7. If it works, it works.

 

[Raelgun]

The life expectancy of a PSU on average is between 5-6 years I believe. Trying to pinpoint the exact time of failure takes us to abstruse nonlinearity way past the deterministic bounds inside chaos and dynamical systems and so it's impossible to answer.

I'm not an expert on PSUs, again, but I believe the cause of most common failures is attributed to degraded capacitors and in case the overvoltage/surge protection fails, other components could be damaged in the process.

Feel free to correct me whenever applicable, always keen on learning.

If you don't mind me asking, where did the 5-6 years expectancy come from?

I am reading on google right now that capacitors last 15-20 years.

 

[Troezar]

I'd say take the warranty period and add a good few years. My AX750 is 7 years old, Corsair would have to assume worst case. Full load high ambient for 7 years? I reckon 10 years at light use will be fine.

 

[Insolent mint]

If you don't mind me asking, where did the 5-6 years expectancy come from?

I am reading on google right now that capacitors last 15-20 years.

From the five year warranty that most reputable manufacturers will grant you (best case scenario). The corsair RM-series being the outlier with a 10-year warranty. Power supplies do not last decades and the capacitors isn't the only variable.

 

[Pulse88]

Just tested 5700xt ref witout frequency (2300 limet) and with 1.35v core. 2310mhz core avg on firestrike Xtreme was possible under cold water. 2365 set in wattman and 910 mem.

Gfx score 14600 for firestrike extreme.

5700xt seems to be an beast under cold water.

 

[Insolent mint]

Just tested 5700xt ref witout frequency (2300 limet) and with 1.35v core. 2310mhz core avg on firestrike Xtreme was possible under cold water. 2365 set in wattman and 910 mem.

Gfx score 14600 for firestrike extreme.

5700xt seems to be an beast under cold water.

But it had already been established that you got a golden chip. Some reviewers only managed to hit 2.05-2.1GHz~ under water.

 

[Raelgun]

From the five year warranty that most reputable manufacturers will grant you (best case scenario). The corsair RM-series being the outlier with a 10-year warranty. Power supplies do not last decades and the capacitors isn't the only variable.

Companies don't put long warranties on products as its not good business for them.

Its also poor misconception that a product only lasts as long as the warranty sticker thats on them and die after.

Warranties for one are put on as law, in the UK its 2 years to start with, anything after then its PR (though good for the consumer), we all know a product will fail but a good design do last and upkeep do last.

 

[Pulse88]

But it had already been established that you got a golden chip. Some reviewers only managed to hit 2.05-2.1GHz~ under water.

Not so sure my chip is golden to early to know still. About reviewers not hitting high clocks on gcn and rdna is no suprice. Many have over the years put highest powertarget and voltage and called it an day. Outcome is throttling and heat and lower avg clocks at different cooling selutions. Those 2.05ghz and 2.1ghz clocks could be no sppt in use or bad paste on gpu when under water. 7nm in both vega vii and 5700xt neads to have great paste when go higher volt and wattage since chip area to cool is smaller. W/mm2 gets higher.

To trust reviewers on Oc is stupid when it comes to Gcn or Rdna. Users gives an better idea what to be expected from cards when it comes to oc.

Pascal and Turing cards is closer to reality when reviewers do their oc under air/stock and water vs real world users oc their cards.

This is the second time you try some kind of dmg controll for you or in Nvidias favor. Last time in this thread you tried to justify 100 pound more for 2070 super vs 5700xt after i posted an voltage scaling table for my card under air and water in firestrike extreme. Now you try to put my sample as golden chip beacuse some reviewers gets lower clocks on normal water.

Anyway here is updated one with chilled water.


Some testing posted earlier in this thread. Some Firestrike extreme gpu score testing 5700xt.

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 1810mhz 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.

chilled water. Softpower table mod. water -13 degrees -8 degress idle gpu

2298/910 mhz 1240mv 270w 2248mhz 50% 30 degrees 14340
2365/910 mhz 1350mv 320w 2310mhz 60% 45 degrees 14598
2370/910 mhz 1350mv 325w 2315mhz 60% 47 degrees fail. card fail in 20 seconds in test 1. To low voltage at given junktion temp.

 

[Voodoo_what]

Please remind us why all the reviewers are wrong.

 

[RedDeangelo]

Please remind us why all the reviewers are wrong.

They are reviewers not overclockers

 

[TrixX]

They are reviewers not overclockers

Steve and Steve both managed to prove they had sod all idea with Vega OC when they OC'd them originally and their subsequent testing. Navi OC is just as fiddly as Vega, probably moreso. HWU was particularly egregious as he maintained he knew what he was doing when he just bombed the card and got crap results.

 

[howkey]

Here is quick info after two hours of testing. I tested the card in the 3DMark Timespy only:
- default clocks are 1750Mhz on a core without drops, undervolting without effect (clocks, temps and power consumption)
- core temp 74°C, +- 1550rpm - > card is not completely quiet, it is slightly audiable, if you wish completely silent card then wait for the Nitro
- power consumption +- 200W card, on a wall +- 310W
- in the Furmark core temp is +- 76°C with same fans rpm
- OC tested in the 3D Mark at [email protected] and 0 power limit, core temp went up to 80°C with +-1550 rpm, consumption +-20W higher, core frequency is +-1920Mhz with peak 2050Mhz
- any change of power limit resulted in failure in the 3D Mark
- any change of a memory frequency resulted in failure in the 3D Mark, but in the Furmark at 950Mhz it ran stable for 20 minutes
- power limit +10 caused core temp hitting 110°C in the Furmark with +-55W higher consumption

 

[R3v4n]

Based on the testing I've done my overall thoughts are -

Yes it OCs quite high in terms of core clock, but performance gain isn't going to change your experience but power draw and temps go up quite a bit.

I feel like the best way to go about it undervolt and even underclock slightly, you get more consistent core speeds, better minimum FPS, better temps, much lower power draw and ofc completely silent.

 

[Orange Nexus]

Based on the testing I've done my overall thoughts are -

Yes it OCs quite high in terms of core clock, but performance gain isn't going to change your experience but power draw and temps go up quite a bit.

I feel like the best way to go about it undervolt and even underclock slightly, you get more consistent core speeds, better minimum FPS, better temps, much lower power draw and ofc completely silent.

On my XT i just tried Auto Undervolt in Wattman and no other changes. Card boost to 2000MHz in Firestrike. Not sure if that was even an oc. But, hot spot temp went from stock 95 to 81c with AUV. Core Temp was 70c. Sustained gaming core temp went as high as 77c and heared the fan a bit taking one headphone on rigth ear.

 

[RedDeangelo]

I can get the anniversary edition for £90 cheaper than the red devil. With the washer mod it could run more quietly, but its still loud.

But at least I'll get it within a week rather than a month.

 

[LeMson]

If AMD reference was a dual axial fan they would be selling like hotcakes, I'd of bought one on launch. My ears over the next 2+ years are worth £90 to me

 

[R3v4n]

If AMD reference was a dual axial fan they would be selling like hotcakes, I'd of bought one on launch. My ears over the next 2+ years are worth £90 to me

I really don't get this unfounded idea people have about the noise lol.

I'm sitting right next to my PC (it's on the desk) any fan speed up to 50% is not audible, and with my current clocks/voltage it sits at 35% while below 80c (which is all the time). Once you get up to 60% you can hear a soft air sound, and ofc it gets exponentionally louder at 100% it is very loud, but there is no need at all for fans to ever be that high lol.

By the comparison the high pitched sound that my Vega 56 Pulse made was considerably worse because that penetrates headphones, the air sound from this card even at 100% does not penetrate headphones.

 

[LeMson]

I really don't get this unfounded idea people have about the noise lol.

I'm sitting right next to my PC (it's on the desk) any fan speed up to 50% is not audible, and with my current clocks/voltage it sits at 35% while below 80c (which is all the time). Once you get up to 60% you can hear a soft air sound, and ofc it gets exponentionally louder at 100% it is very loud, but there is no need at all for fans to ever be that high lol.

By the comparison the high pitched sound that my Vega 56 Pulse made was considerably worse because that penetrates headphones, the air sound from this card even at 100% does not penetrate headphones.

Unfounded?

https://youtu.be/FQJCm7bnOfU?t=1129

https://youtu.be/Q4M8IELhG4s?t=418

Find me some reviews praising the blower cooler?

 

[R3v4n]

Unfounded?

https://youtu.be/FQJCm7bnOfU?t=1129

https://youtu.be/Q4M8IELhG4s?t=418

Find me some reviews praising the blower cooler?

I just gave you first hand experience and I'm not backed/funded/paid/sponsored by any company/brand/manufacturer/agenda like 99.9% of YouTube, it is unfortunately the sad truth in 2019.

The gamers nexus review is honestly complete rubbish I assume his PC is in an oven? I have no clue how he managed to get his temps that high.

I don't doubt that aibs are cooler but it operates well below it's thermal cap either way so you gain nothing other than hwinfo showing you a different number.

Personally I'd definitely avoid Pulse since my Vega 56 Pulse had that obnoxious high pitched sound which is actually super annoying.

But hey do whatever you want it's your money haha .

 

[docsonic]

Not everyone who owns an XT will undervolt & underclock so their experience may be.... this? Yep, 2100rpm on the XT blower is not the jet engine people like to make it out to be.