[WCCFTech]AMD’s FirePro Professional GPUs compared to Quadro counterparts

[NTMBK]

Those mac pro's are good for market share yes, 2 "firepro" gpu's in there, only cheaper, and without the firepro perks.

It's forcing all professional software which runs on Mac to convert from CUDA to OpenCL.

 

[Silverforce11]

Well, that's a very good question. I normally don't like to switch product if I don't have any support issues or performance complaints. It seems that AMD has been boosting their FirePro line lately and it may be worth trying out if my customers are willing. As long as the software has FirePro on the compatibility list.

I normally take the approach of if it's not broke, don't fix it.

This is the key here. Corporations don't like unknowns. They go with what has worked fine in the past.

For AMD to claw into the HPC market, they can only do it by completely beating NV on performance, efficiency and price to attract users to switch. They can't do it by just being competitive.

 

[Hitman928]

Here's another take on the latest Firepro vs Quadro:

AMD is showing some confidence in pricing its FirePro W9100 at $4000. Compared to the slower Quadro K5000 at $1800 and the faster Quadro K6000 at $5000, AMD isn't far off the mark, though. And in the end, the FirePro W9100 surfaces as a strong candidate for high-end workstation duty, particularly when your workload is well-suited to the GPU's strengths (and the driver team's priorities).
How does the FirePro W9100 fare in our final analysis, then? The $4000 card's price tag is justified by excellent performance, versatility across mature professional segments and the latest workloads, and unmatched connectivity. You get a mix of speed in 3D tasks and general-purpose compute-intensive apps, or both at the same time.
It'll be interesting to see how many professionals dig deep for no-compromise speed in their performance-sensitive software. If the audience is out there, AMD's FirePro W9100 should help reclaim some of the company's workstation market share.

Link



Even more so than in the home consumer case, these cards are really case sensitive as far as price/performance goes. There are still some portions of the market AMD has 0 penetration in because of lack of support whether it is CUDA only or AMD's drivers just aren't there yet, but they're improving. As hothardware.com put it:

At the same time, however, AMD has made a strong (and much-improved) argument for itself in a price/performance evaluation. No, it doesn't always match the Quadro K6000 for total performance, but it does often match or exceed that GPU once you factor in the cost of the card. Workstation markets tend to tolerate much higher prices than the consumer space, but compared to where the W9000 was at this point in 2012, the W9100 is a far better product.

Here's what all this means in aggregate: If you need more than 12GB of VRAM, then the W9100 is the obvious best card the market. If you're working in Maya or PTC Creo, the FirePro may be the best all-around choice, period. Other applications are a case-by-case basis -- in many instances, while the FirePro W9100 isn't as fast as the Quadro K6000, its price/performance ratio keep it firmly in the running. If AMD continues to improve its product mix and overall software support, it should close the gap even more in the pro GPU market in the next 18-24 months.

Link 2

 

[ShintaiDK]

This is the key here. Corporations don't like unknowns. They go with what has worked fine in the past.

For AMD to claw into the HPC market, they can only do it by completely beating NV on performance, efficiency and price to attract users to switch. They can't do it by just being competitive.

Corporations dont like lack of support. Thats the key AMD issue is.

 

[Silverforce11]

Corporations dont like lack of support. Thats the key AMD issue is.

Uh, Maya or CAD and a lot of workstation tasks are pretty easy even for dummies. It's not the lack of support (which I have no proof to say yes or no whether that statement is true).

When I think of the major advantage with NV, it is CUDA. With many software/field, AMD isn't an option, period. OpenCL needs a lot more time to mature.

AMD definitely has a mountain to climb, with entrenched CUDA usage & NV brand loyalty along with OpenCL in its infancy.

 

[Genx87]

Performance\dollar is misleading imo. Does the cost of whoever is using the product not count? If you were to add in the cost of the engineer or designer using the hardware + licensing costs for the software. A price\performance shrinks drastically. AMD took the cost of their product and showed the price\performance while leaving out the largest costs in the equation. It may even be if the competitions product is faster when adding in the total cost the competition actually wins price\performance as well.

 

[desprado]

Uh, Maya or CAD and a lot of workstation tasks are pretty easy even for dummies. It's not the lack of support (which I have no proof to say yes or no whether that statement is true).

When I think of the major advantage with NV, it is CUDA. With many software/field, AMD isn't an option, period. OpenCL needs a lot more time to mature.

AMD definitely has a mountain to climb, with entrenched CUDA usage & NV brand loyalty along with OpenCL in its infancy.

Professional are well educated and have much more knowledge than us and there choice is Nvidia as simple it can be explained to all.

 

[mrmt]

Performance\dollar is misleading imo. Does the cost of whoever is using the product not count? If you were to add in the cost of the engineer or designer using the hardware + licensing costs for the software. A price\performance shrinks drastically.

This. If AMD is talking performance/dollar, they are chasing the bottom end of the professional market.

 

[Dribble]

Uh, Maya or CAD and a lot of workstation tasks are pretty easy even for dummies. It's not the lack of support (which I have no proof to say yes or no whether that statement is true).

When I think of the major advantage with NV, it is CUDA. With many software/field, AMD isn't an option, period. OpenCL needs a lot more time to mature.

AMD definitely has a mountain to climb, with entrenched CUDA usage & NV brand loyalty along with OpenCL in its infancy.

All the devs writing all those packages code using nvidia hardware because of the support - nvidia respond to bug reports, actually fix bugs, etc. That means the software tends to work on nvidia, or is at least likely to get patched if they have bugs. On AMD the drivers aren't as good, the dev support isn't as good, bugs don't get fixed as fast, and the devs aren't using the hardware anyway so find less of the problems (perhaps only QA will have the odd AMD box).

Hence as a customer you buy nvidia because it works. That matters significantly more then any upfront hardware cost saving. For AMD to fix this they need to have support as good as nvidia, and this simply isn't the case. JH famously said many years ago they had more software developers then hardware, it's a major part of what nvidia does. For AMD it's always been an afterthought, something other people do and it shows. AMD the cpu company in particular has always left it to other people to do the software which was ok for x86 as MS and Intel wrote it all (although they lost out a bit in compilers as Intel wrote them).

For gpu's that's a big problem as they don't have Intel and MS to do the coding, but as a company AMD just doesn't prioritize software and has failed to adjust. Hence they tend to release a bunch of hardware with new features and then hope the open source community will magically fill in the gaps, and it pretty well always fails.

 

[AtenRa]

On AMD the drivers aren't as good, the dev support isn't as good, bugs don't get fixed as fast, and the devs aren't using the hardware anyway so find less of the problems (perhaps only QA will have the odd AMD box).

Im sure you can provide a proof for those claims ??

 

[RussianSensation]

http://www.xbitlabs.com/articles/graphics/display/nvidia-quadro-k5200-k4200-k2200_11.html

"The new cards are also superior to AMD’s products of the same price categories, offering the most attractive price/performance ratio to professional users."

This seems to go directly against the graph in the OP here.
From the conclusion of the article you linked to. There are more gems in there as well.

How did you miss the part that you have to connect the 50% sale prices posted in AMD's marketing materials with the benchmarks linked at Xbitlabs? Now you get the full picture of the price/performance dominance posted in AMD's slides. Instead you completely ignored why the FirePro at 50% off is such an amazing incentive for professionals to upgrade by using price/performance based on prices in the Xbitlabs' article.

Im sure you can provide a proof for those claims ??

AMD has been eroding NV's professional GPU market share for 4 straight years now. If the drivers and support was so poor, customers would try AMD and forget it, but that's not what we are seeing. Instead we are seeing more and more customers switching and sticking with FirePro. AMD has a very sound strategy by making FirePro a compute and a workstation card in one. That's a major flexibility Quadro can't bring to the table with its crippled DP performance. The interesting part will be seeing how AMD will be able to scale VRAM on FirePro's based on 300 series is they choose HBM.

MacPro update should give us an indication if Apple is really serious about OpenCL or if AMD simply got the design win because Apple thought Tahiti was a better professional chip than GK104 at that time.

 

[kawi6rr]

Prove them wrong then... I bet you can't but decided to come here to troll as usual

:thumbsup: This!!

 

[96Firebird]

AMD has been eroding NV's professional GPU market share for 4 straight years now. If the drivers and support was so poor, customers would try AMD and forget it, but that's not what we are seeing. Instead we are seeing more and more customers switching and sticking with FirePro. AMD has a very sound strategy by making FirePro a compute and a workstation card in one. That's a major flexibility Quadro can't bring to the table with its crippled DP performance. The interesting part will be seeing how AMD will be able to scale VRAM on FirePro's based on 300 series is they choose HBM.

I'm interested in seeing the marketshare numbers for the professional graphics segment, got a link?

 

[sontin]

I'm interested in seeing the marketshare numbers for the professional graphics segment, got a link?

More interessing is the Dollar share. Market share says nothing.
And we see that nVidia's GPU business is growing while AMD GPU business is shrinking.

 

[Keysplayr]

How did you miss the part that you have to connect the 50% sale prices posted in AMD's marketing materials with the benchmarks linked at Xbitlabs? Now you get the full picture of the price/performance dominance posted in AMD's slides. Instead you completely ignored why the FirePro at 50% off is such an amazing incentive for professionals to upgrade by using price/performance based on prices in the Xbitlabs' article.



AMD has been eroding NV's professional GPU market share for 4 straight years now. If the drivers and support was so poor, customers would try AMD and forget it, but that's not what we are seeing. Instead we are seeing more and more customers switching and sticking with FirePro. AMD has a very sound strategy by making FirePro a compute and a workstation card in one. That's a major flexibility Quadro can't bring to the table with its crippled DP performance. The interesting part will be seeing how AMD will be able to scale VRAM on FirePro's based on 300 series is they choose HBM.

MacPro update should give us an indication if Apple is really serious about OpenCL or if AMD simply got the design win because Apple thought Tahiti was a better professional chip than GK104 at that time.

There is a whole lot more to consider than performance per dollar. Dribble's post was actually quite good at describing it.
As 96Firebird had asked, got any links for market share numbers for professional GPUs? Over the last 4 years?

 

[rainy]

I'm interested in seeing the marketshare numbers for the professional graphics segment, got a link?

I couldn't find results for Q3 - those are from Q2.

AMD's share of professional graphics hardware shipments climbs again, on the back of FirePro shipments to Apple


Q2'14 marked the third quarter of shipments for AMD FirePro professional graphics products in Apple's recently revamped Mac Pro, and FirePro numbers continued to climb as Apple's premier professional platform ramped. With 74% of units shipped bearing its Quadro brand, Nvidia still commands the lion's share of the market, but AMD took another few points of market share in the quarter to rise to 25.0%.

http://jonpeddie.com/press-releases...breaks-out-to-record-setting-quarter-in-q214/

 

[96Firebird]

I couldn't find results for Q3 - those are from Q2.


http://jonpeddie.com/press-releases...breaks-out-to-record-setting-quarter-in-q214/

Thanks, but I am more interested in a long-term chart that shows marketshare for each quarter. Want to compare changes with new product releases and see if there is any correlation.

 

[Silverforce11]

I never thought AMD's FirePro would make a dent against NV's dominance in this market, so its rather surprising.

 

[Cookie Monster]

I never thought AMD's FirePro would make a dent against NV's dominance in this market, so its rather surprising.

It could be due to their partnership with Apple.

From what I know, AMD is purely chosen because of their better pricing. They simply offer better performing products given each price bracket. Many small/medium sized businesses that require workstation GPUs may go down the Firepro path since Quadros are much more expensive.

Having said that, there are some instances of driver bugs/glitches with Firepros (just experienced one a few weeks ago). Employees might suffer downtime due to getting these sorted out. Quadros typically don't have any issues and the lack of downtime is less. Downtimes cost alot of money so at the end of the day, the prices could actually even out.

 

[AnandThenMan]

Having said that, there are some instances of driver bugs/glitches with Firepros (just experienced one a few weeks ago). Employees might suffer downtime due to getting these sorted out. Quadros typically don't have any issues and the lack of downtime is less. Downtimes cost alot of money so at the end of the day, the prices could actually even out.

You have data to back this up no doubt.

 

[RussianSensation]

I'm interested in seeing the marketshare numbers for the professional graphics segment, got a link?

"Nevertheless, competitive pricing and improving hardware have helped increase AMD’s professional graphics market share from 12% in 2009 to 20% at present. Strong performance in professional graphics was an important factor that drove AMD’s growth in the last few quarters. In Q1 2014, the company marked its seventh consecutive quarter of revenue and share growth in the professional graphics business segment. AMD is increasing its investment in the area to build a stronger relationship with key customers."
http://www.forbes.com/sites/greatsp...olio-to-increase-its-footprint-in-the-market/

This article was as of June 2014 but it sighted older market share from Q1 2014. By Q3, 2014, AMD was near 25%:

"In 2013, Jon Peddie Research estimated that ~4.9 million professional graphics processors were sold. In 2013, AMD shipped 0.98 million professional GPUs. It commanded 20% of the market. According to Jon Peddie Research, AMD’s competitive pricing and improved hardware helped it increase its market share in the professional graphics market from 12% in 2009 to 25% in 2014." - October 6, 2014
http://finance.yahoo.com/news/amd-professional-graphics-attain-milestone-193408277.html

Some forecast AMD finish Q4 2014 at near 30% but it's tough to say since NV did refresh their lineup in the Fall to make it more competitive.

 

[RussianSensation]

There is a whole lot more to consider than performance per dollar. Dribble's post was actually quite good at describing it.
As 96Firebird had asked, got any links for market share numbers for professional GPUs? Over the last 4 years?

If you want market share for each of 2009, 2010, 2012, 2013, 2014, you know you could Google that yourself right?

http://www.anandtech.com/show/6137/the-amd-firepro-w9000-w8000-review-part-1/3

Market share was at a low of 12% in 2009/2010 and as of last quarter was 25%.

Thanks, but I am more interested in a long-term chart that shows marketshare for each quarter. Want to compare changes with new product releases and see if there is any correlation.

Are you seriously asking us to link you 23 data points (Q1 2009 to Q3 2014)?

That's very hard to find. You either need to be an analyst/associate covering this space at an Equity Research firm which pays for these reports, have connections in the industry or spend time googling press releases, and reading quarterly and annual reports of both NV and AMD and looking for market share comments in there.

 

[RussianSensation]

I never thought AMD's FirePro would make a dent against NV's dominance in this market, so its rather surprising.

A lot of people keep ingoring my point that FirePro doesn't make a distinction between a compute and a workstation card. On the other hand Quadro is worthless for compute which means you need to go out and buy a Tesla to complement it.

http://www.xbitlabs.com/articles/graphics/display/nvidia-quadro-k5200-k4200-k2200_9.html#sect1

All of a sudden you are looking at double the cost or more going with NV if you need that sort of capability. Also, NV purposely crippling Quadro and creating a separate line of essentially the same product with enabled DP under Tesla marketing goes to show the type of firm they are because AMD doesn't ask you to bend over twice since it's the same chip!

 

[AtenRa]

It is funny because it was NVIDIA who was focused on a single GPU for Compute and Gaming up to Fermi. Then they changed that with Kepler and used two different architectures for their professional GPU line up.

And now AMD starting from Tahiti (HD7970) with GCN, are focusing on Compute using a single architecture for both Compute and Gaming much like NVIDIA in the past.

 

[Cookie Monster]

You have data to back this up no doubt.

I don't have scientific data but this is what I have experienced in my field along with what everyone practically has said based on their experience.

Performance is one thing. Its stability that is the most important and bugs can really drive people nuts when deadlines loom.