XFX RX 480 GTR Performance Results

[Kanan]

All of these cards can hit 1500 with proper cooling, bumped power limits and some core voltage. I'm running 1465/2250 benchmarking and 1435/2250 games on the stock bios/core voltage only running out on power limits.

Watch the video and tell me that again. His RX 480 is way better than yours or any other I've seen so far. I had my reasons to call it "golden" obviously.

 

[cdawall]

Watch the video and tell me that again. His RX 480 is way better than yours or any other I've seen so far. I had my reasons to call it "golden" obviously.

He hit 1475? Yes temperatures are lower, but that was a known fact when I purchased mine. Tell you what I bumped the cards fans up and see if it will run 1475 with the lower temps.

Yes the cooler is better on that card, but it isn't exactly ground breaking, the card itself isn't a better PCB. once temps get up the Sapphire and Asus are the same/better.

 

[Recon-UK]

Wow. You're actually arguing that its ok to put a ddc pump/res on top of your gpu. How lazy/negligent does one have to be to argue this point?

 

[thesmokingman]

^^lol

He hit 1475? Yes temperatures are lower, but that was a known fact when I purchased mine. Tell you what I bumped the cards fans up and see if it will run 1475 with the lower temps.

Yes the cooler is better on that card, but it isn't exactly ground breaking, the card itself isn't a better PCB. once temps get up the Sapphire and Asus are the same/better.

The pcb is arguably better, it uses the higher end fets like other recent Amd high end cards but hmm that may be splitting hairs as the nitro+ is at the same level imo. Maybe just a nod to the pcb with the metal ir fets in the xfx and his roaring. Those two share the same pcb so it may be a shared design with Amd.

Watch the video and tell me that again. His RX 480 is way better than yours or any other I've seen so far. I had my reasons to call it "golden" obviously.

I find it sorta ironic that you're making this silly point using someone elses card that we know very little about. cdawall's card is hitting 10mhz under the review sample card that's gone thru who knows how much binning, and 10mhz under at higher temps. Do not see the "way better" here.

 

[cdawall]

The pcb is arguably better, it uses the higher end fets like other recent Amd high end cards but hmm that may be splitting hairs as the nitro+ is at the same level imo. Maybe just a nod to the pcb with the metal ir fets in the xfx and his roaring. Those two share the same pcb so it may be a shared design with Amd.

To be fair the AMD design has a very good PCB setup. The Nitro+ would be second only to the Asus had Sapphire not removed one whole phase since it makes no difference. The XFX is a great design my point is simply that it doesn't matter.

I find it sorta ironic that you're making this silly point using someone elses card that we know very little about. cdawall's card is hitting 10mhz under the review sample card that's gone thru who knows how much binning, and 10mhz under at higher temps. Do not see the "way better" here.

Watched the video again and noticed a peculiar 36mv voltage bump so in short screw his card. This was a 30mv bump which puts me at 1.168v under load vs his 1.175, with fans kicked up temps stayed under 65C for the run.

 

[Ungari]

Adored TV has come out with a compelling explanation of JayTwoCents benchmarks of both the Gigabyte Fail, and the XFX GTR Golden Chip being due to binning of the chips by TDP which explains the wide variance in temperatures and OC headroom:

Particularly intriguing is the speculation that had the process been better giving more high-end yields that we could have seen an RX 490 come out of Polaris.
I think Adored may be correct in guessing that next year will have a Polaris refresh as the Global Foundries improves it's process, much the same as Nvidia has done with Maxwell, Maxwell 2.0, and Pascal(Maxwell 3.0).

 

[cdawall]

Particularly intriguing is the speculation that had the process been better giving more high-end yields that we could have seen an RX 490 come out of Polaris.
I think Adored may be correct in guessing that next year will have a Polaris refresh as the Global Foundries improves it's process, much the same as Nvidia has done with Maxwell, Maxwell 2.0, and Pascal(Maxwell 3.0).

I need to test this on my brothers XFX, but for perspective I am pulling 150w (stock 1342mhz) to 175w (+50% power and 1435mhz). It would be interesting to see if this card is drawing more due to temps or due to ASIC quality.

 

[Durvelle27]

Would using a 6 Pin to 8 Pin adapter limit OC potential

 

[cdawall]

Would using a 6 Pin to 8 Pin adapter limit OC potential

It shouldn't you would just draw more from the 6 pin wires

 

[Durvelle27]

It shouldn't you would just draw more from the 6 pin wires

I only ask because I can not get 1400Mhz and even at 50% Target it never breaks 168W

 

[cdawall]

I only ask because I can not get 1400Mhz and even at 50% Target it never breaks 168W

Could be a BIOS limit on wattage, or a weak GPU my brother has just the regular XXX edition with the short PCB and is hitting 1435 on 20%

 

[Onglar]

got a gtr black edition coming on Monday, and i know its gonna clock well.... counting on self fulfilling prophecies here

 

[Ungari]

Would using a 6 Pin to 8 Pin adapter limit OC potential

My understanding was that a dedicated 8-pin with it's sensor, would deliver more consistent power to the card without increasing draw from the PCI-E slot, making for smoother delivery to maintain the OC.
However, this would not be an issue if you were experiencing similar low power usage as JayTwoCents XFX GTR card.
Is the reason for the adapter your PSU?

 

[cdawall]

My understanding was that a dedicated 8-pin with it's sensor, would deliver more consistent power to the card without increasing draw from the PCI-E slot, making for smoother delivery to maintain the OC.
However, this would not be an issue if you were experiencing similar low power usage as JayTwoCents XFX GTR card.
Is the reason for the adapter your PSU?

The card wont know the difference between a true 8 pin and a 6 pin with an 8 pin adapter. Quite a few powersupplies already double the grounds over for an 8 pin anyway this just adds a connector to do exactly that.

 

[Durvelle27]

My understanding was that a dedicated 8-pin with it's sensor, would deliver more consistent power to the card without increasing draw from the PCI-E slot, making for smoother delivery to maintain the OC.
However, this would not be an issue if you were experiencing similar low power usage as JayTwoCents XFX GTR card.
Is the reason for the adapter your PSU?

no reason just used it