Vega Rumors

Ieldra said:
An 899$ P4000 outperforms Vega.

You claimed this comparison was not accurate because Vega is not using pro drivers like quadro. I will post a screenshot of your post when I'm home :)
Click to expand...
Certified ones, which is the pro use case, it isn't. At least in the case of Linux, pro and gaming are one in the same though. Windows drivers likely the same. Actual pro drivers will be different with the certification and support.

You can't seriously think P4000 is anywhere close to Vega in gaming performance?

8GB vs 16GB
5.3 vs 12.5 FP32 TFLOPs
Actually does async

Should I include FP16 performance where P4000 doesn't even quote a figure? Isn't that 25 vs 0.33 TFLOPs?
 
Anarchist4000 said:
At least in the case of Linux, pro and gaming are one in the same though.
Click to expand...
In case of Windows, there is a well documented disrepancy. Hell, Wx7100 beats Vega FE in those benchmarks, are you happy?
Anarchist4000 said:
Should I include FP16 performance where P4000 doesn't even quote a figure? Isn't that 25 vs 0.33 TFLOPs?
Click to expand...
Do quote it because i want to laugh at how silly you will look with your stupid numbers when gaming benchmarks hit and instead of being 75 times faster as you are peddling here, it is not even 40% faster (over 1070).
Anarchist4000 said:
Actually does async
Click to expand...
lol
 
lolfail9001 said:
In case of Windows, there is a well documented disrepancy. Hell, Wx7100 beats Vega FE in those benchmarks, are you happy?

Do quote it because i want to laugh at how silly you will look with your stupid numbers when gaming benchmarks hit and instead of being 75 times faster as you are peddling here, it is not even 40% faster (over 1070).

lol
Click to expand...
The 330gflops of fp16 are for packed math, SM6.1 (Gp102, GP104, GP106) has one vec2 fp16 unit per SM so as you can imagine it's only used by sensible developers who don't subscribe to the absurd view that you should just use FP32 math at that point. Nonsense.

I am fully aware that Vega FE is far more powerful than 25% cut GP104 core, but it performs just like it in the standard benchmark for cad/cam (spec view) according to AMD's own internal testing. This is not something debatable, we have all seen the performance because I posted it earlier.

My point has always been that Vega FE doesn't seem even remotely competitive for cad/cam, and I expect it will underwhelm in games as well. It makes no sense whatsoever to market this card for professionals( with the appropriate drivers!!!!) and compare it to a gaming card with gaming drivers, unless they are trying to avoid comparison to actual competition.

For machine learning applications Vega offers very competitive performance *on paper* but their rocm etc are not well supported yet, whereas Cuda and Cuda accelerated libraries have been around for a while.

What exactly is the purpose of Vega FE? Machine learning? Mi25 right?

There's going to be another Vega pro later this year with certification from solidworks autodesk etc, gaming Vega is supposedly releasing next month...

This seems like it only exists so they can say they met their deadline of H1 17 honestly
 
Ieldra said:
This seems like it only exists so they can say they met their deadline of H1 17 honestly
Click to expand...

Yep

And the lawsuits will launch anyhow unless they somehow meet the magical efficiency gains they promised. Doesn't seem they will. Sucks to be Raj and this isn't even the piglet he wanted to show for the fall fair, this is someones else' piglet.

Not that making it Intel/RTG will fix that.. it seems the design is less than spectacular. I smell Bulldozer.
 
  • Like
Reactions: N4CR
like this
Simplyfun said:
The ones they promised over their own last gen. Keep up.
Click to expand...
Funny you say that, because the only efficiency gains over last gen they promised were with Polaris. They did not deliver on those either, but they never promised them for Vega.
 
lolfail9001 said:
Funny you say that, because the only efficiency gains over last gen they promised were with Polaris. They did not deliver on those either, but they never promised them for Vega.
Click to expand...

The charts say different. But hey, whatever works for you.

Twenty bucks says there's a class action over Vega mentioning efficiency down the road. BTC or cash. Let's not split hairs and just put the coins with a trusted agent.
 
Ieldra said:
My point has always been that Vega FE doesn't seem even remotely competitive for cad/cam, and I expect it will underwhelm in games as well. It makes no sense whatsoever to market this card for professionals( with the appropriate drivers!!!!) and compare it to a gaming card with gaming drivers, unless they are trying to avoid comparison to actual competition.
Click to expand...
Budget conscious prosumers purchasing workstations with minimal graphics and replacing them with Titans and consumer drivers was the target AMD stated. How is the comparison invalid as that's exactly what AMD claims consumers are doing already? This card really shouldn't be used for CAD and content creation.

Ieldra said:
For machine learning applications Vega offers very competitive performance *on paper* but their rocm etc are not well supported yet, whereas Cuda and Cuda accelerated libraries have been around for a while.

What exactly is the purpose of Vega FE? Machine learning? Mi25 right?
Click to expand...
The purpose is to enable application developers. Hence no certified drivers and just hardware. I keep saying that, but you keep missing that point. The card is intended as a development kit where bugs and crashes are to be expected. Not caked into a server for deep learning. The extent of deep learning use should be to write the code to implement it.

Ieldra said:
There's going to be another Vega pro later this year with certification from solidworks autodesk etc, gaming Vega is supposedly releasing next month...
Click to expand...
There's also a dual Vega10 in documentation as a pro card to accompany it.

lolfail9001 said:
I know, after all nobody stops one from turning fp16 into fp32 and then back.
Click to expand...
For whatever reason people get mad when you compare FP16 to FP32 performance because peak flop rates aren't comparable or something.
 
Anarchist4000 said:
Budget conscious prosumers purchasing workstations with minimal graphics and replacing them with Titans and consumer drivers was the target AMD stated. How is the comparison invalid as that's exactly what AMD claims consumers are doing already? This card really shouldn't be used for CAD and content creation.


The purpose is to enable application developers. Hence no certified drivers and just hardware. I keep saying that, but you keep missing that point. The card is intended as a development kit where bugs and crashes are to be expected. Not caked into a server for deep learning. The extent of deep learning use should be to write the code to implement it.


There's also a dual Vega10 in documentation as a pro card to accompany it.


For whatever reason people get mad when you compare FP16 to FP32 performance because peak flop rates aren't comparable or something.
Click to expand...

The alledged people buying Titans instead of Quadro cards are the same people who will buy RX Vega consumer cards instead of Vega FE cards... You can't have it both ways, people are smart enough to buy professional cards or they aren't.
 
Last edited:
Anarchist4000 said:
Budget conscious prosumers purchasing workstations with minimal graphics and replacing them with Titans and consumer drivers was the target AMD stated. How is the comparison invalid as that's exactly what AMD claims consumers are doing already? This card really shouldn't be used for CAD and content creation.
Click to expand...


err not really, last generation pro cards which will be heavily discounted once the next gen come out, beat the pants of any of these cards in a pro situation. So its not even smart to look at Vega FE or latest Titan for pro uses. At least not for the things they showed. It is a good card if you want some pro performance without the cost and that is it.

The purpose is to enable application developers. Hence no certified drivers and just hardware. I keep saying that, but you keep missing that point. The card is intended as a development kit where bugs and crashes are to be expected. Not caked into a server for deep learning. The extent of deep learning use should be to write the code to implement it.
Click to expand...

Application development, possibly, but AMD has traditionally been weak on the pro side of applications anyways. And price has never been major factor in selling pro cards. Either the performance is there or they won't be used.

There's also a dual Vega10 in documentation as a pro card to accompany it.
Click to expand...

And did it sell well?


For whatever reason people get mad when you compare FP16 to FP32 performance because peak flop rates aren't comparable or something.
Click to expand...

They are not comparable, two totally different markets. Outside of DL, FP 16 isn't used anywhere too much yet, and Vega isn't about to change that. Just for comparative measures, why not take FX series FP 16 vs AMD's 24 bit for the 9700 pro? Would that have been fair at that time? Were developers swayed to make a nV specific FP16 path over AMD's 24 bit full precision path? I didn't see it happen did.... I did see nV's path being dumbed down to DX8 though. Its all about the money at the end for developers. Having new features is great, but if AMD is selling a product that needs a specific path to make it work as good as the competition, forget it, its not going to fly, not with their 30% marketshare, with a product that is for 5% of that 30%.
 
Last edited:
In the end, what makes the FE Vega 10 any different then Vega 10 cards for gaming? Same chip, same HBM except less of it for the game cards. Why would AMD say the card is not meant for gaming? If I was a a game developer - making content would this card not be ideal? Anyways this card looks to be more for the semi-professional crowd who has an option to use Pro drivers or gaming drivers, rich students maybe as well. I wonder if AMD will have student discounts?

No warm fuzzy feeling here but will hold out to see what RTG delivers. To me it is clear RTG comparing the same price point but using Pro Drivers tweaked for certain applications against the Titan Xp shows that it probably can't compete against it in games. The 1080Ti performance is not much different then the Titan Xp - so what does that tell everyone? I hope either RTG is sand bagging to surprise us or just have some rather good performance/$ cards.

Also the 300w rating - lol - big card - there goes the smaller package size advantage for HBM. I have to say the card design looks awesome! Puts the Nvidia jagged, disjointed, trying to be cool Bat-mobile style to shame, with simple to the point form follows function design.
 
Last edited:
Ocellaris said:
The alledged people buying Titans instead of Quadro cards are the same people who will buy RX Vega consumer cards instead of Vega FE cards... You can't have it both ways, people are smart enough to buy professional cards or they aren't.
Click to expand...
There's still a difference in memory capacity, which is significant. So that's not having it both ways. For professionals with large datasets capacity would be significant and amplified by HBCC.

lolfail9001 said:
Nope, simply FP16 is kind of irrelevant outside of compute and mobile games.
Click to expand...
Or games with HDR, particles, hair, lots of simple physics, etc that make up a significant amount of work. Stuff found in high end and enthusiast gaming.

noko said:
In the end, what makes the FE Vega 10 any different then Vega 10 cards for gaming? Same chip, same HBM except less of it for the game cards. Why would AMD say the card is not meant for gaming? If I was a a game developer - making content would this card not be ideal? Anyways this card looks to be more for the semi-professional crowd who has an option to use Pro drivers or gaming drivers, rich students maybe as well. I wonder if AMD will have student discounts?
Click to expand...
Well clockspeeds and price would be significant. Say 15% faster and a third the cost for gamers. The gaming version as Raja said will be faster, most likely from higher clockspeeds.
 
Anarchist4000 said:
There's still a difference in memory capacity, which is significant. So that's not having it both ways. For professionals with large datasets capacity would be significant and amplified by HBCC.
Click to expand...


Operative word there "Could" have to wait and see.

Or games with HDR, particles, hair, lots of simple physics, etc that make up a significant amount of work. Stuff found in high end and enthusiast gaming.
Click to expand...

Well hair and geometery stuff (which is pretty much everything in that list other than HDR), is all going to finally bring them to parity with Pascal's geometry through put. So as I stated, unless its going to give some ridiculous advantage for the developers on everyone's card, doing a separate path just will eat up their bottom line. Maybe AMD will pony up the money to help developers.....

Always comes back down to the almighty $, if it doesn't make financial sense it won't be done.

Well clockspeeds and price would be significant. Say 15% faster and a third the cost for gamers. The gaming version as Raja said will be faster, most likely from higher clockspeeds.
Click to expand...

Nope. Not going to happen. Didn't Fiji or any AMD launch in the past tell us anything? Price performance comparative.

The only one that wasn't was HD4xxx and 5xxx that was because AMD had a die advantage so they could do it.

And again, $ talks BS walks. The pricing should tell us where Vega will end up.
 
razor1 said:
They are not comparable, two totally different markets. Outside of DL, FP 16 isn't used anywhere too much yet, and Vega isn't about to change that. Just for comparative measures, why not take FX series FP 16 vs AMD's 24 bit for the 9700 pro? Would that have been fair at that time? Were developers swayed to make a nV specific FP16 path over AMD's 24 bit full precision path? I didn't see it happen did.... I did see nV's path being dumbed down to DX8 though. Its all about the money at the end for developers. Having new features is great, but if AMD is selling a product that needs a specific path to make it work as good as the competition, forget it, its not going to fly, not with their 30% marketshare, with a product that is for 5% of that 30%.
Click to expand...
You mean besides new consoles? A platform with a specific path that covers the vast majority of the gaming market revenue? So yeah it seems all but guaranteed FP16 will see extensive use. We've already seen the feature used. Devs will just have to spend time dumbing down their code to accommodate Nvidia in the PC market. Not that different from disabling async for Nvidia already.

razor1 said:
And did it sell well?
Click to expand...
Can't say we've seen them used for deep learning or server accelerators yet.
 
Anarchist4000 said:
You mean besides new consoles? A platform with a specific path that covers the vast majority of the gaming market revenue? So yeah it seems all but guaranteed FP16 will see extensive use. We've already seen the feature used. Devs will just have to spend time dumbing down their code to accommodate Nvidia in the PC market. Not that different from disabling async for Nvidia already.
Click to expand...




New consoles, is Vega in new consoles? Did Polaris have extensive FP16 performance?

Hmm WTF? Why the fuck do you make this shit up, I have no damn clue man.

Its much more different than nV, nV has 70% of the market

AGAIN ALMIGHTY $

Can't say we've seen them used for deep learning or server accelerators yet.
Click to expand...

Yeah exactly.
 
Anarchist4000 said:
Budget conscious prosumers purchasing workstations with minimal graphics and replacing them with Titans and consumer drivers was the target AMD stated. How is the comparison invalid as that's exactly what AMD claims consumers are doing already? This card really shouldn't be used for CAD and content creation.
.
Click to expand...
Budget conscious prosumers as shown by AMD focus on CAD related would just go with the cheaper P4000 as Ieldra keeps mentioning.
Look AMD is screwing the narrative up like I explained awhile back on B3D but I have given up now trying to reiterate this; they are either deliberately skewing comparisons or have a confused narrative what the market is for the RX Vega.
If it is meant to be workstation visualisation then they lose out to the cheaper P4000 even though they keep insisting on using the Titan xp, this is even more emphasised that they stated the drivers were not game orientated awhile back....
But they also go on about its FP16 and FP32 compute capabilities, which raises why bother with the Mi25 where this is more important rather than gaming.
Look how Nvidia has managed to perform well in gaming in the past with substantially less compute performance with 980ti against Fury X, yeah there is a need for FP32 and possibly to a lesser or future extent FP16 but both of those are primarily for modelling/scientific-enterprise computation/etc rather than gaming, where they are important but not as important as other factors (results as I mentioned show this with 980ti compared to Fury X).
That is also ignoring FP64 that unfortunately for now still relies upon the Hawaii products until next Vega version 2018 (I think).

So IMO it is a very confused product for now that is trying to be promoted in the workstation professional visualisation but against the wrong Nvidia product, and when they do mention an Nvidia Quadro (as it was indirectly in the PCGamer preview) they totally ignore the lower models and go straight to the GP102 one, which is substantially quicker than the P4000 that for now the RX Vega is struggling against in CAD/visualisation.
On top of this is the Grid/cloud solution that Nvidia is pushing for professional Quadro world and sharing of the compute/graphics capabilities of a node solution between users and/or applications.
IMO this is another framework/platform AMD needs to build up like they are doing for now in HPC.

Cheers
 
Last edited:
Gideon said:
Why are we arguing over the non gaming version of Vega? Do any of us really care about it? Let me know when the gaming version of Vega is here and we can finally end this debate on how it will perform.
Click to expand...

Because AMD chose to benchmark their professional card against a gaming card in professional benchmarks to skew the results. The only benchmarks people have to talk about are from an unfair comparison.
 
Ocellaris said:
Because AMD chose to benchmark their professional card against a gaming card in professional benchmarks to skew the results.
Click to expand...

AMD would never do such a hype ish thing ... so it ain't so.

AMD marketing may be the most useless marketing organization on the face of the planet Earth. The only thing that matters are independent reviews.
 
Gideon said:
Why are we arguing over the non gaming version of Vega? Do any of us really care about it? Let me know when the gaming version of Vega is here and we can finally end this debate on how it will perform.
Click to expand...


I agree with ya but there is nothing else to talk about right now, and reality, all we can expect from the gaming versions of Vega over the FE would be at most 10% differences anyways for games, maybe 15% and for course the base clocks, which will probably be a little bit lower than the FE. AMD typically uses TDP or TBP as their power consumption average as of late. Lets see if they do the same here.
 
Ocellaris said:
Because AMD chose to benchmark their professional card against a gaming card in professional benchmarks to skew the results. The only benchmarks people have to talk about are from an unfair comparison.
Click to expand...

And when ever AMD has done this, the dark side shows up as not being not really being competitive...... Shit they did it with Ryzen too.
 
Ocellaris said:
Because AMD chose to benchmark their professional card against a gaming card in professional benchmarks to skew the results. The only benchmarks people have to talk about are from an unfair comparison.
Click to expand...

Only a fool believes what marketing tells them about a product. Until we have reviews from a independent source we dont have much of anything. Since we know AMD is weak at being a professional card I fail to see how much that tells us about the card in gaming. Thankfully we should know in a month.
 
razor1 said:
http://support.amd.com/en-us/download/frontier?os=Windows+10+-+64

BTW game mode Vega driver available.
Click to expand...
WTF did I just watch??? Measured in Centuries? Is that how long AMD is going to take?

Virtualization - Games - Virtualization - VR and AR. Meaning it better be damn good at gaming if it is good at Virtualization. I hate dog and pony shows. grrrrrrr We just need some independent honest reviews and block our ears from AMD painful marketing. Perf/$ rules in the end. If it performs less then the competition, doesn't matter how much lipstick you put on it - price it right, it will sell.
 
razor1 said:
And when ever AMD has done this, the dark side shows up as not being not really being competitive...... Shit they did it with Ryzen too.
Click to expand...
I do like that you can run Pro drivers on the card if you have specific applications that can take advantage of it. Issue is would one be switching back and forth with drivers if you want a more diversified setup? We just need some solid, down to the point reviews. Also would like to know if AMD VR has been brought up to better standings or works better.
 
Simplyfun said:
AMD would never do such a hype ish thing ... so it ain't so.

AMD marketing may be the most useless marketing organization on the face of the planet Earth. The only thing that matters are independent reviews.
Click to expand...

AMD engineers are frequently able to work magic on relatively low R&D budgets. Sometimes I think we don't give 'em enough credit for that.

But AMD's Marketing moonies are functional retards, and need to be dropped from low Earth orbit. It'd do everyone a lot of good.

Early indications from the usual rumor mill suspects are saying Vega is performing "like a 1080 Ti". But, as some other folks have said, best to just wait for the independent reviews before drawing conclusions.
 
noko said:
I do like that you can run Pro drivers on the card if you have specific applications that can take advantage of it. Issue is would one be switching back and forth with drivers if you want a more diversified setup? We just need some solid, down to the point reviews. Also would like to know if AMD VR has been brought up to better standings or works better.
Click to expand...


Well the thing is pro drivers still don't compensate for the deficit, against true pro cards, its a nice thing to have, yeah ok, I can see that, but really any one serious about professional work, will have a system dedicated for their work and one for their games which will be great for their counterpart apps, not an all in one machine that is ok in both.

So is this card targeted at people that are scrapping by with money (professional)? is that what AMD is trying to say, if a person doesn't have the money to get two systems for their work and gaming, get this? They must be out of work if they can't afford two systems. Its not like they upgrade every year either at least not all components.

There is no such thing as "value" branding in business, they have to be able to do the work and that is why they demand the money. that is why certification is there.

Prosumers "value" is up to the prosumer, a 1200k card that can't compete with a 800 dollar quadro that is a 400 buck difference that right there is a good start for a gaming system.
 
Last edited:
razor1 said:
Well the thing is pro drivers still don't compensate for the defect, against true pro cards, its a nice thing to have, yeah ok, I can see that, but really any one serious about professional work, will have a system dedicated for their work and one for their games which will be great for their counterpart apps, not an all in one machine that is ok in both.

So is this card targeted at people that are scrapping by with money (professional)? is that what AMD is trying to say, if a person doesn't have the money to get two systems for their work and gaming, get this? They must be out of work if they can't afford two systems. Its not like they upgrade every year either at least not all components.
Click to expand...
Well if you make content using professional software such as Adobe Products to make games and play them - Use game engines like in the video using the Unreal engine (game) and then SteamVR, lol, it would be better if AMD just made the drivers do both well or have easy configuration change from one type of work environment to the next. Do agree having specialized rigs is more a mark of a professional plus computers today are cheap compared to 20 years ago costs. Back in 1989 I bought a whopping 20meg (yep meg) Supra hard drive with 2mb (yep mb) of EDO ram for $950 for my Amiga 500. Inflation over that time would make that like $1500 today. So a $1200 Titan Xp or FE Vega for what you get is peanuts compared to back then. We just have to see once folks can start reviewing them to figure out WTH. RTG is so vague with weird off the wall points of almost any discussion with Vega it becomes pointless to even pay attention to RTG.
 
noko said:
Well if you make content using professional software such as Adobe Products to make games and play them - Use game engines like in the video using the Unreal engine (game) and then SteamVR, lol, it would be better if AMD just made the drivers do both well or have easy configuration change from one type of work environment to the next. Do agree having specialized rigs is more a mark of a professional plus computers today are cheap compared to 20 years ago costs. Back in 1989 I bought a whopping 20meg (yep meg) Supra hard drive with 2mb (yep mb) of EDO ram for $950 for my Amiga 500. Inflation over that time would make that like $1500 today. So a $1200 Titan Xp or FE Vega for what you get is peanuts compared to back then. We just have to see once folks can start reviewing them to figure out WTH. RTG is so vague with weird off the wall points of almost any discussion with Vega it becomes pointless to even pay attention to RTG.
Click to expand...

I agree and its never a good sign when marketing is being evasive (also giving more reasons why to buy something, never works well for marketing either). There has to be something there they don't want to share at the moment or there is a weakness.

What I would like to say at this point when looking at Vega FE is Vega FE is a good pro card that can game, nV's Titan is a great gaming card that can do some pro stuff. (the only reason I got titan was because I needed the vram for PBR texture makers, and going to a quadro with more ram just didn't make financial sense, 3d work for what I do at home, titan was more than enough)
 
razor1 said:
I agree and its never a good sign when marketing is being evasive. There has to be something there they don't want to share at the moment or there is a weakness.
Click to expand...
Well that is exactly what I get from RTG so far, evasive, non-answering, told to wait etc.
 
noko said:
Well that is exactly what I get from RTG so far, evasive, non-answering, told to wait etc.
Click to expand...


I think the performance is going to be decent a bit lower than the 1080ti for the most part. but power consumption is going to be much higher. So its going to be a repeat of Fiji for the most part with out the scaling at higher resolutions, due to the shader flop parity between the competitive cards.
 
Last edited:
Well well, didn't see this:
To playtest and optimize the gaming experience, the exclusive ability of the Radeon™ Vega Frontier Edition to switch from Radeon™ Pro Settings to Radeon™ Settings and back with a couple of clicks enables rapid switching between software features for faster iteration during development workflows.

When in “Gaming Mode” the full suite of gaming features of Radeon™ Software are made available, including Radeon™ Chill 5 and Radeon™ WattMan 6.

Game developers can also use the wealth of no-cost open development tools and software found on GPUOpen.com to optimize their next-generation gaming experiences for the pinnacle of AMD graphics technology.
  • I Am Focused on Compute Workloads
    The ROCm driver will be available for users on June 29th.
  • I Am Focused on Workstation Workloads
    If you are a workstation user on a Windows® platform, get your driver here.
    If you are a workstation user on a Linux® platform, get your driver here.
  • I Am Focused on Game Development Workloads
    Radeon™ Vega Frontier Edition introduces “Gaming Mode”. This allows you to switch from Radeon™ Pro Settings to Radeon™ Settings where you can access gaming features for playtesting and performance optimization.
    If you are focused on game development on a Windows® platform, get your driver here and switch to “Gaming Mode.”
  • I Just Want to Game
    If you want to game on a Windows® platform, get your driver here and switch to “Gaming Mode.”
Click to expand...
http://pro.radeon.com/en-us/product/radeon-vega-frontier-edition/

So looks like the drivers will have modes optimized for different work loads. Sounds good if they deliver. Maybe I should read more but usually end up hating it :(.
 
CSI_PC said:
Look how Nvidia has managed to perform well in gaming in the past with substantially less compute performance with 980ti against Fury X, yeah there is a need for FP32
Click to expand...
Tiled rasterization reducing the workload 15-40%? That one software feature accounts for nearly all their lead in poorly optimized titles.

CSI_PC said:
If it is meant to be workstation visualisation then they lose out to the cheaper P4000 even though they keep insisting on using the Titan xp
Click to expand...
It's a development card. Any pro creating models or using CAD shouldn't be using this card. Maybe devs making CAD applications. Engine development or experimentation with any GPU acceleration then sure. Artists, outside indie devs, have other options. Best analogy is that FE looks like a XBox development kit.

CSI_PC said:
yeah there is a need for FP32 and possibly to a lesser or future extent FP16 but both of those are primarily for modelling/scientific-enterprise computation/etc rather than gaming,
Click to expand...
Maybe on Nvidia products where they lack the capability. Game devs have already started using FP16 on console. Next generation of engines furthering that use on PC. Based on all the people I've seen arguing against it, I'm guessing consumer Volta will lack it as well.

noko said:
Do agree having specialized rigs is more a mark of a professional plus computers today are cheap compared to 20 years ago costs.
Click to expand...
Indie devs with third party engines and low budgets are more of a thing now.
 
Anarchist4000 said:
Maybe on Nvidia products where they lack the capability. Game devs have already started using FP16 on console. Next generation of engines furthering that use on PC. Based on all the people I've seen arguing against it, I'm guessing consumer Volta will lack it as well.
Click to expand...

not on the PC side. The only console that has FP16 is PS4 pro and dev's are not using much of FP16 at the moment even on those because the performance increase of using FP16 on it doesn't yield much more performance. PS4 pro can only barely play 4k games even with FP16 optimizations, and that doesn't bode well for future games. So looking at PS4pro vs Xbox Scorpio which doesn't have FP16 optimizations, you already are at 50% of the console market, and factor in the PC market, yeah, still no go on FP16 utilization for a year or two.

Indie devs with third party engines and low budgets are more of a thing now.
Click to expand...

Most companies lease quadro cards. There is no need to pay for full price on these cards for businesses.

Now that is a different market than the Prosumer at home...... So don't know how we can parallelize those two groups. The marketing doesn't make sense at this point, is it prosumer, is it indie developers that are actual companies, what? leases for 1 year the company only pays for about half the cost of the quadros.
 
Last edited:
You must log in or register to reply here.
Back
Top