Wow, really?

[rich1051414]

So, apparently I just topped the FireStrike leader board for users with an 8320 and a single R9 290. I find that a bit unbelievable as I only use an NH-D14 , and an Asus DirectCUII with a Kraken G10 and AIO mod applied(Zalman LQ320), plus memory and vrm heat sinks of course.

Here is a link to the score

CPU is actually running at 5Ghz, 23x215. 
GPU running at 1240 GPU Clock, 6800 effective memory clock, 1412 mV, 150% power target. Max Temp after 3 FireStrike runs is 52C. 





Edit: Sorry I should have posted this in the 3dmark competition thread.

Edit: Found a new stable clock speed, I may try increasing this score later today if I have time.

Not much of a difference though, I fear i am at my limit pushing the cpu any further, stability is getting too hard to locate.

Edited August 2, 2014 by rich1051414

[ccokeman]

Very nice! Thats one hell of a clock for a 290 on air! 

[rich1051414]

Very nice! Thats one hell of a clock for a 290 on air! 

Actually the 290 water cooled. Well, sorta, an all in one. That alone seems a bit overkill though, since I am limited by the max limit of the voltage, and not heat. 

Edited August 2, 2014 by rich1051414

[ccokeman]

Even so its still pretty stout. 

[rich1051414]

Even so its still pretty stout. 

While monitoring the gpu voltage from a third party tool, it is showing it is using 1.22-1.32. However, it is set to use 1.412, is this massive inconsistancy a normality? Would something be starving the card of voltage, is the sensor just way off, or is something else going on here?

[ccokeman]

Voltage droop occurs during peak loads. The other way to verify voltage is with a multimeter so you get a true picture of the voltage that is applied.  

[rich1051414]

Voltage droop occurs during peak loads. The other way to verify voltage is with a multimeter so you get a true picture of the voltage that is applied.  

Well, the voltages everywhere else are spot on, if this were the case, wouldn't I be able to see drops in multiple locations, especially on the CPU which is drawing quite a bit more than the GPU? I was assuming the VRM's were just not delivering on their voltage, but I was hoping there was something simple to correct that.

My PSU is 3 weeks old and its a NZXT HALE82 700W. It shouldn't break a sweat.

Edited August 2, 2014 by rich1051414

[ccokeman]

Voltage droop is a function of the VRM and design. By using a multimeter you validate that you are actually supplying the applied 1.412v at idle so you can tell if you are indeed seeing a droop or if the applied voltage of 1.412v is just indicated rather than actually applied. You can see this when you extend the overclocking limits.   

[rich1051414]

Voltage droop is a function of the VRM and design. By using a multimeter you validate that you are actually supplying the applied 1.412v at idle so you can tell if you are indeed seeing a droop or if the applied voltage of 1.412v is just indicated rather than actually applied. You can see this when you extend the overclocking limits.   

Ah, I gotcha. I appreciate the knowledge That makes complete sense.