Overclock.net › Forums › Graphics Cards › AMD/ATI › [Official] AMD R9 Radeon FURY / NANO / X / Pro DUO FIJI Owners Club
New Posts  All Forums:Forum Nav:

[Official] AMD R9 Radeon FURY / NANO / X / Pro DUO FIJI Owners Club - Page 929

post #9281 of 11553
Oops, sorry frown.gif I didn't read that.
post #9282 of 11553
Quote:
Originally Posted by LionS7 View Post

@Krzych04650

My R9 Fury X is very sad. I was testing around 2 weeks, cos it crashes not too often. After crash on 1.24V, witch is +42mV, I put 1.26V... and, I dont know anymore. This is for 1100Mhz core. You need to test these cards in games like Star Wars Battlefront, FrostBite 3 i killing them. Try it. Cos I don't think, that I need 1.26V, 1100Mhz for other games...

Overclocking is not really a thing with those cards, at least from what I see on my Fury. I tested this 1100 MHz -48mV more extensively and it crashes and needs stock voltage, and any overclock above that will crash even with full overvoltage. And for only 50MHz less at 1050 MHz I can set voltage to -90 mV and be stable. Gains from overclocking compared to power draw and temperature increase are just...pathetic.

As for FrostBite3, I think that Witcher 3 is also very merciless in terms of overclocking and stability, I always had to set my cards lower than for other games to be stable. I was testing FreeSync for like an hour running around Novigrad and some woods and swamps and it never crashed at -90mV stock or 1100 MHz stock voltage.
Quote:
Originally Posted by Spartoi View Post


What did you use to measure GPU wattage/voltage? I use HWInfo and my Fury Tri-X is runs at 1080mhz/545mhz with a -12mV undervolt. HWInfo reports that after 10 minutes of Valley Extreme HD, my GPU uses about 271W and 1.181V. Considering you're stock undervolt reading, I would think you'd use less wattage and voltage than my card. Does the Nitro use more power than the Tri-X?

I don't know about Tri-X, but Tri-X is reference PCB and Nitro is custom, so there may be differences. This would explain temperature and noise difference between the two in tests.

I am using HWInfo and MSI Afterburner for voltage monitoring. As for wattage, I am just measuring total power draw from the wall with wall meter or however it is called in English. And this includes monitor and speakers because I was too lazy to reconnect it, so you can cut ~25W from those measurements.
Quote:
Originally Posted by AngryLobster View Post

Man I have no idea how you manage 61-62C.

Ambient here is 22C and copying your settings (1440P, same undervolt, 35% fixed fan speed, etc.), mine hits 75C and sits there or bouncing up to 76. Same game too (Witcher 3). This is with a open side panel too.

I don't know what is causing a almost 15C different in temps we are seeing but either something is wrong with my card or you live at the North Pole.
Quote:
Originally Posted by AngryLobster View Post

As I said above, side panel is open. I'm going to try applying new thermal paste right now in case the factory job is really bad.

EDIT: Yeah just changed paste, made zero different in temperatures.

Toms required around 1450RPM to maintain 75C in their review. My experience matches theirs when I close my side panel. I gotta stop believing everything on the internet.

Tests are made for ambient temperature of 23,7-24,1 degrees. I am measuring at night because during the day I have around 28C ambient temperature. Basically very hot room with 2 big windows and no ventilation, and only 2-3 degrees cooler than outside, so if there is like 36C outside than you can imagine that I am dying there biggrin.gif

My case is Fractal Design Define R5. I am using slow RPM Silentium PC Sigma Pro 140mm fans set to 5V, which gives 400 RPM. I was afraid that they will be useless since you can barely feel any air pushed by them if you place your hand behind one, but actually they work well. There are two for front intake, one for bottom intake (1-2 degrees improvement for GPU), one for side intake (lowers temperature by another 4-5 degrees) and rear exhaust. Those measurements for fans, how much each helps, was made with R9 390, I didn't test them yet for Fury, but temp goes up quickly if I turn case fans off, so there is some airflow after all even with those not very well performing in theory fans.

Keep in mind that -90mV undervolt lowered temps from 71C to 61-62 C. On stock voltage I am getting 71.

This is how the temp scales with fan speed after undervolting (10 minutes Heaven Benchmark).:



I can make video of 10 minutes Valley run, what recording software doesn't take 500 GB for 10 minutes and won't take video on 50 parts like FRAPS?
Edited by Krzych04650 - 7/28/16 at 2:48am
post #9283 of 11553
I don't know much sense does it make, probably not much, but I also downclocked Fury to 390 performance that I also tested recently and measured power draw and temperatures.


This probably is quite pointless, but it shows that you can scale Fury to whatever performance you want and power draw will scale very well with it. Downclocking to 390 performance (around 25% performance decrease) and setting -50% power target surely draws much less power than stock card at ~75% load.

post #9284 of 11553
Quote:
Originally Posted by Krzych04650 View Post

Overclocking is not really a thing with those cards, at least from what I see on my Fury. I tested this 1100 MHz -48mV more extensively and it crashes and needs stock voltage, and any overclock above that will crash even with full overvoltage. And for only 50MHz less at 1050 MHz I can set voltage to -90 mV and be stable. Gains from overclocking compared to power draw and temperature increase are just...pathetic.

As for FrostBite3, I think that Witcher 3 is also very merciless in terms of overclocking and stability, I always had to set my cards lower than for other games to be stable. I was testing FreeSync for like an hour running around Novigrad and some woods and swamps and it never crashed at -90mV stock or 1100 MHz stock voltage.
I don't know about Tri-X, but Tri-X is reference PCB and Nitro is custom, so there may be differences. This would explain temperature and noise difference between the two in tests.

I am using HWInfo and MSI Afterburner for voltage monitoring. As for wattage, I am just measuring total power draw from the wall with wall meter or however it is called in English. And this includes monitor and speakers because I was too lazy to reconnect it, so you can cut ~25W from those measurements.

Tests are made for ambient temperature of 23,7-24,1 degrees. I am measuring at night because during the day I have around 28C ambient temperature. Basically very hot room with 2 big windows and no ventilation, and only 2-3 degrees cooler than outside, so if there is like 36C outside than you can imagine that I am dying there biggrin.gif

My case is Fractal Design Define R5. I am using slow RPM Silentium PC Sigma Pro 140mm fans set to 5V, which gives 400 RPM. I was afraid that they will be useless since you can barely feel any air pushed by them if you place your hand behind one, but actually they work well. There are two for front intake, one for bottom intake (1-2 degrees improvement for GPU), one for side intake (lowers temperature by another 4-5 degrees) and rear exhaust. Those measurements for fans, how much each helps, was made with R9 390, I didn't test them yet for Fury, but temp goes up quickly if I turn case fans off, so there is some airflow after all even with those not very well performing in theory fans.

Keep in mind that -90mV undervolt lowered temps from 71C to 61-62 C. On stock voltage I am getting 71.

This is how the temp scales with fan speed after undervolting (10 minutes Heaven Benchmark).:



I can make video of 10 minutes Valley run, what recording software doesn't take 500 GB for 10 minutes and won't take video on 50 parts like FRAPS?

Thanks for the information. My friend and I both at -90mv are seeing results 10-12C higher. I don't really care for Valley or Heaven. Witcher 3 at 4K VSR brings the card close to 80C. At 1440P with your settings myself and my friend are both just under 75C.

I'm not sure what the discrepancy is given you are using fans at 400RPM which are essentially doing nothing in a case that large but 62C is literally impossible for me unless ambient was 7C.
post #9285 of 11553
Hey all so I finally got my hands on a working Z170 motherboard and decided that I would like to try 4K on my single Fury X. I have been tunnel vision focused on 2560 x 1440 144hz panels for the past few years so I am looking for a good 4K monitor to game on and if it has Freesync with a range that is manageable on a single Fury X that is all the better. Looking for suggestions from people with experience with 4K panels. I don't mind a TN if the color quality is decent. As far as panel size I am willing to consider anything from 28 or higher up to the max size as long as it is a monitor not a TV. Cheers! biggrin.gif
Kung Fury X
(16 items)
 
  
CPUMotherboardGraphicsRAM
i7 6700k MSI Z170A XPower Gaming Titanium  Radeon R9 Fury X  16GB (8x2GB) Corsair Vengeance DDR4 
Hard DriveHard DriveOptical DriveCooling
500GB Samsung 850 EVO 1TB WD Black Some external ASUS garbage Custom EK PETG with Bitspower Fittings 
OSMonitorKeyboardPower
Win10 Pro 64 bit BenQ XL2730Z 2560 x 1440 @ 144hz Logitech G710+ Blue EVGA SuperNOVA P2 750 Watt 
CaseMouseAudioOther
NZXT 440 (White) Logitech g502 Bose Companion 2 Series III Satellite Speakers Xbox One Elite Controller 
  hide details  
Reply
Kung Fury X
(16 items)
 
  
CPUMotherboardGraphicsRAM
i7 6700k MSI Z170A XPower Gaming Titanium  Radeon R9 Fury X  16GB (8x2GB) Corsair Vengeance DDR4 
Hard DriveHard DriveOptical DriveCooling
500GB Samsung 850 EVO 1TB WD Black Some external ASUS garbage Custom EK PETG with Bitspower Fittings 
OSMonitorKeyboardPower
Win10 Pro 64 bit BenQ XL2730Z 2560 x 1440 @ 144hz Logitech G710+ Blue EVGA SuperNOVA P2 750 Watt 
CaseMouseAudioOther
NZXT 440 (White) Logitech g502 Bose Companion 2 Series III Satellite Speakers Xbox One Elite Controller 
  hide details  
Reply
post #9286 of 11553
Quote:
Originally Posted by Thoth420 View Post

Hey all so I finally got my hands on a working Z170 motherboard and decided that I would like to try 4K on my single Fury X. I have been tunnel vision focused on 2560 x 1440 144hz panels for the past few years so I am looking for a good 4K monitor to game on and if it has Freesync with a range that is manageable on a single Fury X that is all the better. Looking for suggestions from people with experience with 4K panels. I don't mind a TN if the color quality is decent. As far as panel size I am willing to consider anything from 28 or higher up to the max size as long as it is a monitor not a TV. Cheers! biggrin.gif

From what I searched, all 4K FreeSync panels has 40-60Hz FreeSync range, except for one, that is iiyama G-Master GB2888UHSU-B1 Gold Phoenix, it has 35-60. It is a TN.
Quote:
Originally Posted by AngryLobster View Post

Thanks for the information. My friend and I both at -90mv are seeing results 10-12C higher. I don't really care for Valley or Heaven. Witcher 3 at 4K VSR brings the card close to 80C. At 1440P with your settings myself and my friend are both just under 75C.

I'm not sure what the discrepancy is given you are using fans at 400RPM which are essentially doing nothing in a case that large but 62C is literally impossible for me unless ambient was 7C.

I don't know why there are such huge temperature differences, especially if like you said my setup is definitely not set for amazing cooling performance, it is set for maximum silence. I never got 80C on this card, even on crazy test like Furmark which makes GPU draw so much more than games I was getting 77C with 35% fan speed and 71C with 50% fan speed. But in games there is no way for such high temps with undervoltage. After overvoltage and 1125 OC I was indeed getting 78C, but this was 120mV above my undervolted setting.
Edited by Krzych04650 - 7/28/16 at 2:08pm
post #9287 of 11553
If no one mentioned it yet, there is a way to hack the monitor down to 32 refresh freesynch.
post #9288 of 11553
Quote:
Originally Posted by dagget3450 View Post

If no one mentioned it yet, there is a way to hack the monitor down to 32 refresh freesynch.

Yea there is but no always works. I wasn't able to make it work properly on my 34UC98, but FreeSync on my unit is not working properly even at stock 55-75 and creates some stutters, flickering and some kind of "micro breaks" in GPU operation, so I am getting refund. But anyways, there are a lot of people who didn't manage to hack FreeSync range.
post #9289 of 11553
New official driver kids. 16.7.3

release notes.

http://support.amd.com/en-us/kb-articles/Pages/Radeon-Software-Crimson-Edition-16.7.2-Release-Notes.aspx



Cheers

Edited.
Edited by bluezone - 7/28/16 at 4:33pm
post #9290 of 11553
Quote:
Originally Posted by bluezone View Post

New official driver kids. 16.7.3

release notes.

http://support.amd.com/en-us/kb-articles/Pages/Radeon-Software-Crimson-Edition-16.7.2-Release-Notes.aspx



Cheers

Edited.
You posted the notes for 16.7.2 wink.gif Here's the actual notes: http://support.amd.com/en-us/kb-articles/Pages/Radeon-Software-Crimson-Edition-16-7-3-Release-Notes.aspx
New Posts  All Forums:Forum Nav:
  Return Home
  Back to Forum: AMD/ATI
Overclock.net › Forums › Graphics Cards › AMD/ATI › [Official] AMD R9 Radeon FURY / NANO / X / Pro DUO FIJI Owners Club