Here's the big test I promised.AMD Sapphire R9 280X (Vapor-X dual fan) Windows 7 x64 2x 24" Display @1080p
Using: sgminer-v5-Lyra2RE-RC1-WIN32 (default Lyra2RE kernel)
PC Specs: Phenom II X6 1100t @4GHz, 16GB RAM (@1600MHz), Samsung EVO 500GB SSD (system), peripherals and network - all wired AMD driver
13.12 - skipped for now. Please provide your experience.
AMD driver
14.4 WHQL best:
Test length: 3 reboots, a minimum of 30-minute-run after each
Memory 1500, Clock 1030, Powertune 20, Intensity 15
900+KHps 70°C (VRM 68°C/81°C) 1760RPM HW:0 TDP-VRM:85W
UV: 66°C (VRM 66°C/76°C) 1640RPM HW:0 TDP-VRM:56W
YouTube 1080p/60fps video plays fine with occasional stuttering.
Display count (1-2) did only make minor difference regarding hash rate.
Photoshop does not bug out the driver nor make it fail.
Cakewalk Sonar X3 - 80+ tracks, spec-intensive project loaded, played and worked great on low-latency. No visual glitches.
Minecraft is a bit choppy but playable and does not bug out the driver nor make it fail. 50+ FPS most of the time.
Stonehearth is unplayable. Don't Starve is good.
Launching Chrome (x64, 10 extensions) with 13 pages works flawlessly (scrypt-n fails at being smooth big time). Stays responsive and fluent while dragging between screens.
~~ Intensity 16-20 yields 910-940+KHps but with significantly more stuttering and occasional mouse and screen lag (no HW errors, though).
Conclusion: This driver is great overall, very stable, good performance.
AMD driver
14.6 RC2 best(?):
Test length: 3 reboots, ran until driver failed
Memory 1500, Clock 1030, Powertune 20, Intensity 15
865+KHps 69°C (VRM 71°C/80°C) 1760RPM HW:5 (test1) TDP-VRM:77-86W (fluctuating)
UV: 66°C (VRM ?°C/?°C) ?RPM HW:? TDP-VRM:xxx
YouTube 1080p/60fps video - could not conduct test.
Display count (1-2) - did not check difference.
Photoshop - did not conduct test.
Games - did not conduct test.
Launching Chrome - did not conduct test.
Cakewalk Sonar X3 - single track project's visuals made the driver fail.
Conclusion: Driver failed during all tests. Unstable.
AMD driver
14.7 RC1 best:
Test length: 3 reboots, a minimum of 30-minute-run after each
Memory 1500, Clock 1030, Powertune 20, Intensity 15
850+KHps 69°C (VRM 70°C/79°C) 1800RPM HW:0 TDP-VRM:76-86W (fluctuating)
UV: 69°C (VRM 69°C/77°C) 1600RPM
HW:1 TDP-VRM:56-62W (fluctuating)
Whilst undervolted, driver did not tolerate intensity 16 and above on the longer run.
YouTube 1080p/60fps video plays fine with occasional stuttering.
Display count (1 vs. 2) made a 30+KHps difference regarding hash rate.
Photoshop does not bug out the driver nor make it fail.
Cakewalk Sonar X3 - 80+ tracks, spec-intensive project loaded, played and worked great on low-latency. No visual glitches.
Minecraft is very choppy nearly unplayable but does not bug out the driver nor make it fail. 25- FPS most of the time.
Stonehearth is unplayable. Don't Starve is decent.
Launching Chrome (x64, 10 extensions) with 13 pages works, stays responsive when dragging between screens.
~~ Intensity 16-20 yields 860-870+KHps but with significantly more stuttering and occasional mouse and screen lag (a few HW errors time to time).
Conclusion: This driver is an interesting one. The more I rebooted the system, the worse the hash rate got. It seems to be stable, though. Will investigate further.
AMD driver
14.9 skipped until further improvements:
I tried it before. Memory frequency, clock and voltage gets locked permanently.
Final conclusion and thoughts:14.4 WHQL worked the best for me.
Not everyone has dedicated mining rigs. The purpose of my test is to see what level of usability my PC maintains when mining Vertcoin-Lyra2RE. I was aiming to achieve a balance between the highest hash rate and the best PC performance.
The results of this test may not apply to all AMD R9 280X cards and may vary on different systems. This should be taken as mere guidance only.That is also the reason I was not using "xintensity 64" during the tests. Tested clock frequencies: 1000-1100MHz (with an increment of 10).
Clock frequency @1020, @1050, @1080 all worked well but changing the clock did not make big a difference.
One more thing. I really enjoyed this whole testing and I'm very happy with what I've seen. It is amazing that we will have an algo that doesn't render my computer useless and annoying, draws way less power, silent and fan-friendly.
The last question is: Will I turn on the radiator this winter?
