You have mentioned this fact multiple times, but the only professional benches. I only seem to recall the initial pcper review, were it was a wash (either XP or Quadro being faster, sometimes significantly so). Have I missed some benches were VEGA wins Titan XP in all those workloads?Vega FE beats the Titan XP in workstation applications, computing, AI, VR, etc...
Great post.Decal filtering has nothing to do with anti aliasing.
Anisotripic filtering has nothing to do with antialiasing.
8x anisotropic has no effect on performance and is simply just default as part of ultra settings.
Also AMD those framerates can be very similar to a gtx 1080 cards.
Yeah had that thought as well in a previous post. "There is no bad publicity"-Strategy. But let's be honest, that will only work if RX Vega is significantly faster and that FE now and FE will then also benefit from this new driver. And with significantly faster I mean at least 20%.
If there is only a minor increase and they price it low it's kind of meh. Not really interesting. We can all say power use doesn't matter but in fact it does a bit. My room is easily 5°C+ warmer after an hour of gaming and currently in summer with outside temps around 30C (86F) and no AC it kind of matters.
Honestly I'd wait for RX to come out to judge it. It seems odd that AMD would claim higher IPC over the last generation as recently as like a month ago only to have Vega have a severe IPC regression compared to two generations ago.Someone ran benchmarks with the Vega FE clocks matched to those of Fury X (including adjusting HBM clocks to make up for the narrower bus width on Vega). The result was that Vega FE did 16% worse, clock-for-clock, than Fury X.
Either the currently released drivers are alpha-level crap, or this is a Bulldozer-class failure.
Someone ran benchmarks with the Vega FE clocks matched to those of Fury X (including adjusting HBM clocks to make up for the narrower bus width on Vega). The result was that Vega FE did 16% worse, clock-for-clock, than Fury X.
Either the currently released drivers are alpha-level crap, or this is a Bulldozer-class failure.
Honestly I'd wait for RX to come out to judge it. It seems odd that AMD would claim higher IPC over the last generation as recently as like a month ago only to have Vega have a severe IPC regression compared to two generations ago.
These kind of tests are only relevant if they are done in exact same machine and Windows (and other software) by the same tester. There are just too many random factors that make that kind of test irrelevant.Someone ran benchmarks with the Vega FE clocks matched to those of Fury X (including adjusting HBM clocks to make up for the narrower bus width on Vega). The result was that Vega FE did 16% worse, clock-for-clock, than Fury X.
Either the currently released drivers are alpha-level crap, or this is a Bulldozer-class failure.
Agree. As i remember this stuff from years ago research shows you get best results if you manage to get consumers to have high expectation yet exceed those expectations. Like eg: It will be faster even though it ought (should) not quite to be so fast. "Will" and "should" is then managed as different types of expectations.
It kinds of fits this story. If vega is as you say plus 20% faster than fe they get this effect. Drama and turmoil and exceeding the "should" expectations because fe set the bar low.
Now. If - IF- rx is plus 20% faster. Lets say 30% we get in this situation. Then we will see the result. If the effect is as the theories and research they will have made a gigantic stunt against what we beliewe. Congrats. Mind you my knowledge is 23 years old here.
Imo this is very risky marketing business. If rx is not at least plus 10% faster in gaming it will come back to their face so hard its incredible damaging. Then they will stand as amateurs not even trying to manage expectations. I find it hard to beliewe as 390 p10 and rebrand launch was handled fairly well considering what they had to compete against. But its amd business so you never know.
Which smelled like?All this smells so much the same as the 2900XT...
Even if RX Vega comes with drivers that boost gaming performance to 1080Ti levels, the 300W power draw is still a problem for me.
That's a huge amount of power to use and a huge amount of heat that must be vented by my case.
The result was that Vega FE did 16% worse, clock-for-clock, than Fury X.
They are basically running a Fiji GPU with a CPU with an IPC of 30-40% above the Vega setup. Fair? I think not
Even if RX Vega comes with drivers that boost gaming performance to 1080Ti levels, the 300W power draw is still a problem for me.
That's a huge amount of power to use and a huge amount of heat that must be vented by my case.
That being said, we don't know if RX Vega will target the same 300W TDP as Vega FE. Or the frankly ridiculous 375W of the Vega FE liquid cooled edition.
The question is not if it is fair, but if it is relevant. In this particular case IPC difference of 40% is irrelevant because the test is GPU limited.
It does not help your case to post Cinebench numbers either.
If we exclude for a moment, that Vega is worse than Fiji clock-for-clock this would indicate some big issues with the driver - which currently is the most likely interpretation of the result.
With pretty much any other benchmark, I would agree with you, but Superposition doesn't stress the CPU at all. Even my ancient 2500K (soon to be upgraded to 1700X btw) was below 50% usage the whole time..
The result in that reddit thread is a Fury X coupled with a 8-core Ryzen CPU.
They thought it was ok to run a Vega with a server CPU from 2010. The server CPU is a 6 core too. Atleast they tried to clock them the same, but the IPC gap is so big between them.
Its actually not irellevant if it chokes the GPU.
No GPU operates on its own and a 7 year old CPU with a poor IPC for sure have the potential of greatly influencing the performance of a big GPU like Vega
I wouldnt trust the results before we see a test with the same setup (reviewers, coff coff)
I guess that the point is that custom VEGA probably use even more than the ref. Like custom Pascals do. Thus the difference stays but custom VEGAs just use even more power?What? 1080Ti bone stock uses more than 250 watts. Overclocked it eclipses 300 watts easy.
What? 1080Ti bone stock uses more than 250 watts. Overclocked it eclipses 300 watts easy.
Well. My first impression and post was also this is insane launch (the post was removed in a cleaning process lol).With today's information at the speed of light world, this is just asking for negative repercussions. Maybe in the old days when controlled environments were privy to info and hardware for benching/leaking/teasing, today you got engineers/creators on Twitter boasting and promoting their own products. And when finalized products don't meet expectations, the backlash is ridiculous. For a recent good example look at No Man Sky.
Even more so than before, today your first impression is most likely your lasting impression. Hawaii never recovered from the launch reviews, 2-3 years later those reviews were still being shared because most consumers don't really dig. "290X review" into Google and you'd get day one launch reviews, not revised reviews with better coolers.
This is why the Vega FE launch irritates me so much. Sure, AMD could have an ace up their sleeve (buy basing this on prior GPU launches, I doubt it). But AMD still went head first with this marketing cluster screw that they've most likely already lost sales on RX Vega. And the last time something this similar happened the usual AMD people shrugged it off as "those aren't very much sales" but they're still sales - that AMD needs.
One bullet point that AMD's PR team just had to ring the bell on basically destroyed the perception of RX Vega.