Vega FE benchmarked by random dude

[PCGuy_5960]

Just now, Valentyn said:

Stop with the logic! I wanted a new blue toy!

PAINT YOUR 980 TI BLUE!

[Valentyn]

Just now, PCGuy_5960 said:

PAINT YOUR 980 TI BLUE!

 

Yeah, but the driver's control panel won't also be blue

[Sacriledgy]

Just now, Valentyn said:

 

Yeah, but the driver's control panel won't also be blue

Put Intel HD graphic driver overlay on it 

 

Profit.

[PCGuy_5960]

Just now, Valentyn said:

Yeah, but the driver's control panel won't also be blue

Just spray paint the control panel blue /s

[PCGuy_5960]

I really hope that Vega can beat the 1080 Ti by at least 10%, because otherwise that 300+W TDP will be hard to swallow

[Valentyn]

4 minutes ago, PCGuy_5960 said:

I really hope that Vega can beat the 1080 Ti by at least 10%, because otherwise that 300+W TDP will be hard to swallow

 

Yeah man, also looking at the dodgy firestrike scores. "IF" true, it means Vega regressed in IPC nearly 30% over Fiji when comparing performance per clock speed.

I find that rather hard to believe; although given all the issues this "reviewer" chap had; maybe he just messed up it all.

[NumLock21]

12 minutes ago, PCGuy_5960 said:

I really hope that Vega can beat the 1080 Ti by at least 10%, because otherwise that 300+W TDP will be hard to swallow

Did you even look at the OP?

[PCGuy_5960]

Just now, NumLock21 said:

Did you even look at the OP?

Yep, but newer drivers may improve performance

[Valentyn]

Just now, PCGuy_5960 said:

Yep, but newer drivers may improve performance

Or you know, correct ones.

[NumLock21]

1 minute ago, PCGuy_5960 said:

Yep, but newer drivers may improve performance

The reviewer had a under power system for the card. I'm surprise that freaking thing even turned on. Your basing your conclusion on that poor and crappy invalid test. Be optimistic, dagnabbit!

[XenosTech]

57 minutes ago, NumLock21 said:

The reviewer had a under power system for the card. I'm surprise that freaking thing even turned on. Your basing your conclusion on that poor and crappy invalid test. Be optimistic, dagnabbit!

But honestly though if he bought this FE card how the heck doesn't he have a better psu or have the common sense to see he's using the wrong drivers too ???

Edited June 28, 2017 by wkdpaul
Cleaned up

[imreloadin]

15 minutes ago, Valentyn said:

Or you know, correct ones.

Shhh you risk exposing the OP's agenda if you try to bring logic into this thread, as a matter of fact it'll get your post deleted

[NumLock21]

23 minutes ago, XenosTech said:

well seeing who posted the article.... Logic and optimism goes right out the window... But honestly though if he bought this FE card how the heck doesn't he have a better psu or have the common sense to see he's using the wrong drivers too ???

This is what happens when you watch a "video" that says high wattage power supplies are stupid and totally believe in it. Not only that, those that actually believe in that nonsense went out in full force, bashing those who bought high wattage power supplies, as if they know everything, while in reality they know nothing.

 

How to buy a psu

1. Set budget

2. Look at reputable brands

3. Look at PSU ratings you like to have

4. Buy the one with the highest wattage within your budget.

5. Give 2 middle fingers to those who makes fun of you for getting a 1K+ watt psu

 

 

 

 

[MageTank]

Why do I keep seeing this "under spec PSU" argument being thrown around? I count 2 8 pin connectors on the card. If they are following ATX spec, that means the card can safely pull 375w. His CPU was a 4790k. Considering he wasn't throwing a power virus at it, you can realistically expect 80-100w from that CPU under normal loads depending on clock speed. That still leaves a 70-100w buffer for the rest of the system. I've run overclocked 980 Ti's on a 500w PSU before, and that's with a 6700k and 32gb of extremely OC'd ram. Depending on quality of PSU (something nobody seems to have information on aside from it's wattage) you can push higher than it's rated capacity as long as it's cool enough.

 

The driver argument is one that will hold, but I highly doubt he ran into power constraints just because he used a 550w PSU. If that is truly the case, AMD has more explaining to do for breaking ATX specs once again. 

[HKZeroFive]

Wow, people are more optimistic than I thought. Nothing is truly set in stone until the actual reviews come out but there are signs that say Vega isn't looking as good as it should be.

 

1. The launch of this Vega FE card has been extremely quiet. No build up. No hype. No samples for reviewers to test. It's different from AMD's usual MO and all it does is imply that AMD is hiding something.

2. The CES demo was terrible in itself. Screen tearing galore.

3. Are people seriously blaming the 550W power supply? If a unit cannot supply enough power, the whole system or program crashes... it doesn't throttle a specific component (which is the card in this case). It's more likely a case of thermal throttling thanks to the power-hungry nature of the card.

4. Drivers. The last line of hope, as always. The release of the "consumer" card is about a month away and drivers aren't optimised yet? Don't kid yourself. The guy even used the "gaming" profile during his test run. And let's be real. Drivers will probably only provide a 5% boost, which isn't going to get this card to GTX 1080Ti performance levels anytime soon if this is true.

 

I truly hope I'm wrong, and again, nothing's set in stone until the reviews emerge. But man, y'all need to start killing the hype and just hope it isn't as bad as it looks to be currently.

[cj09beira]

14 minutes ago, MageTank said:

Why do I keep seeing this "under spec PSU" argument being thrown around? I count 2 8 pin connectors on the card. If they are following ATX spec, that means the card can safely pull 375w. His CPU was a 4790k. Considering he wasn't throwing a power virus at it, you can realistically expect 80-100w from that CPU under normal loads depending on clock speed. That still leaves a 70-100w buffer for the rest of the system. I've run overclocked 980 Ti's on a 500w PSU before, and that's with a 6700k and 32gb of extremely OC'd ram. Depending on quality of PSU (something nobody seems to have information on aside from it's wattage) you can push higher than it's rated capacity as long as it's cool enough.

 

The driver argument is one that will hold, but I highly doubt he ran into power constraints just because he used a 550w PSU. If that is truly the case, AMD has more explaining to do for breaking ATX specs once again. 

agreed, but he did say the card's frequency was all over the place, (thermal or power issues)

[MageTank]

Just now, cj09beira said:

agreed, but he did say the card's frequency was all over the place, (thermal or power issues)

Yeah, could also be firmware issues with the card itself. We saw the same thing with the RX 480 launch if you remember.

[Sacriledgy]

1 minute ago, HKZeroFive said:

Wow, people are more optimistic than I thought. Nothing is truly set in stone until the actual reviews come out but there are signs that say Vega isn't looking as good as it should be.

 

1. The launch of this Vega FE card has been extremely quiet. No build up. No hype. No samples for reviewers to test. It's different from AMD's usual MO and all it does is imply that AMD is hiding something.

2. The CES demo was terrible in itself. Screen tearing galore.

3. Are people seriously blaming the 550W power supply? If a unit cannot supply enough power, the whole system or program crashes... it doesn't throttle a specific component (which is the card in this case). It's more likely a case of thermal throttling thanks to the power-hungry nature of the card.

4. Drivers. The last line of hope, as always. The release of the "consumer" card is about a month away and drivers aren't optimised yet? Don't kid yourself. The guy even used the "gaming" profile during his test run. And let's be real. Drivers will probably only provide a 5% boost, which isn't going to get this card to GTX 1080Ti performance levels anytime soon if this is true.

 

I truly hope I'm wrong, and again, nothing's set in stone until the reviews emerge. But man, y'all need to start killing the hype and just hope it isn't as bad as it looks to be currently.

Agreed, as i said I'm extremely worried about RX Vega performance

[cj09beira]

Just now, MageTank said:

Yeah, could also be firmware issues with the card itself. We saw the same thing with the RX 480 launch if you remember.

wasn't that power issues from too much vcore? (might be wrong though)

[MageTank]

Just now, cj09beira said:

wasn't that power issues from too much vcore? (might be wrong though)

Yeah, and it was causing the cards to run hotter (and boost lower). We might be seeing the exact same thing this time around. AMD did fix it with drivers/firmware though, so it could very well be that kind of issue again.

[cj09beira]

Just now, MageTank said:

Yeah, and it was causing the cards to run hotter (and boost lower). We might be seeing the exact same thing this time around. AMD did fix it with drivers/firmware though, so it could very well be that kind of issue again.

dam they need a new way to set vcore for each individual card so that the cards run cooler fast

[Arc_Jester]

1 minute ago, cj09beira said:

dam they need a new way to set vcore for each individual card so that the cards run cooler fast

 

2 minutes ago, MageTank said:

Yeah, and it was causing the cards to run hotter (and boost lower). We might be seeing the exact same thing this time around. AMD did fix it with drivers/firmware though, so it could very well be that kind of issue again.

Again I'd like to point out that the driver used does not support Vega and is 6 months old.

[MageTank]

Just now, Arc_Jester said:

 

Again I'd like to point out that the driver used does not support Vega and is 6 months old.

Yeah, and again i'd point out that driver/firmware issues are the only logical defense for this cards performance. People need to stop blaming the PSU being used, because 550w is plenty enough for the tests he used. If need be, I can demonstrate what a 5.2ghz 7700k draws under 3dmark load. Hint: Under Prime95 it only pulls 120w, lol. 

[cj09beira]

2 minutes ago, Arc_Jester said:

 

Again I'd like to point out that the driver used does not support Vega and is 6 months old.

[porina]

1 minute ago, Arc_Jester said:

Again I'd like to point out that the driver used does not support Vega and is 6 months old.

Check out the "featured comment" on the videocards page. It seems to suggest the long version number is current, and that the displayed Crimson version in GPU-Z is incorrect.