Friday, November 14th 2014
Big Swing in Market Share From AMD to NVIDIA: JPR
Jon Peddie Research (JPR), the industry's research and consulting firm for graphics and multimedia, today announced estimated graphics chip shipments and suppliers' market share for 2014 2Q in its Market Watch quarterly PC graphics report, an industry reference since 1988.
Graphics processors, stand-alone discrete devices, and embedded processor-based GPUs are ubiquitous and essential components in all systems and devices today -- from handheld mobile devices, PCs, and workstations, to TVs, servers, vehicle systems, signage, game consoles, medical equipment, and wearables. New technologies and semiconductor manufacturing processes are taking advantage of the ability of GPU power to scale. The GPU drives the screen of every device we encounter -- it is the human-machine interface.
The third quarter is typically the big growth quarter, and after the turmoil of the recession, it appears that trends are following the typical seasonality cycles of the past.
Quick report highlights:
GPUs are traditionally a leading indicator of the PC market, since a GPU goes into every system before it is shipped, and most of the vendors are guiding cautiously for Q4 '14.
The Gaming PC segment, where higher-end GPUs are used, was a bright spot in the market in Q3. Nvidia's new high-end Maxwell GPUs sales were strong, lifting the ASPs for the discrete GPU market.
Q3 2014 saw a flattening in tablet sales from the first decline in sales last quarter. The CAGR for total PC graphics from 2014 to 2017 is up to almost 3%. We expect the total shipments of graphics chips in 2017 to be 510 million units. In 2013, 454 million GPUs were shipped and the forecast for 2014 is 468 million.
The quarter in general
Graphics chips (GPUs) and chips with graphics (IGPs, APUs, and EPGs) are a leading indicator for the PC market. At least one and often two GPUs are present in every PC shipped. It can take the form of a discrete chip, a GPU integrated in the chipset or embedded in the CPU. The average has grown from 1.2 GPUs per PC in 2001 to almost 1.55 GPUs per PC.
For PC and mobile device related companies small and large, new to the industry or established, it is critical to get a proper grip on this highly complex technology and understand its future direction. In this detailed 50-page data-based report, JPR provides all the data, analysis and insight needed to clearly understand where this technology is today and where it's headed. This fact and data-based report does not pull any punches: frankly, some of the analysis and insight may prove to be shocking.
Findings include discrete and integrated graphics (CPU and chipset) for Desktops, Notebooks (and Netbooks), and PC-based commercial (i.e., POS) and industrial/scientific and embedded. This report does not include the x86 game consoles, handhelds (i.e., mobile phones), x86 Servers or ARM-based Tablets (i.e. iPad and Android-based Tablets), or ARM-based Servers. It does include x86-based tablets, Chromebooks, and embedded systems.
Graphics processors, stand-alone discrete devices, and embedded processor-based GPUs are ubiquitous and essential components in all systems and devices today -- from handheld mobile devices, PCs, and workstations, to TVs, servers, vehicle systems, signage, game consoles, medical equipment, and wearables. New technologies and semiconductor manufacturing processes are taking advantage of the ability of GPU power to scale. The GPU drives the screen of every device we encounter -- it is the human-machine interface.
The third quarter is typically the big growth quarter, and after the turmoil of the recession, it appears that trends are following the typical seasonality cycles of the past.
Quick report highlights:
- AMD's overall unit shipments decreased 7% quarter-to-quarter, Intel's total shipments increased 11.6% from last quarter, and Nvidia's jumped 12.9%.
- The attach rate of GPUs (includes integrated and discrete GPUs) to PCs, for the quarter was 155% (up 2%) and 32% of PCs had discrete GPUs, (flat from last quarter), which means 68% of PCs today are using the embedded graphics in the CPU.
- The overall PC market increased 6.9% quarter-to-quarter, and decreased 2.6% year-to-year.
- Desktop graphics add-in boards (AIBs) that use discrete GPUs increased 7.8% from last quarter.
GPUs are traditionally a leading indicator of the PC market, since a GPU goes into every system before it is shipped, and most of the vendors are guiding cautiously for Q4 '14.
The Gaming PC segment, where higher-end GPUs are used, was a bright spot in the market in Q3. Nvidia's new high-end Maxwell GPUs sales were strong, lifting the ASPs for the discrete GPU market.
Q3 2014 saw a flattening in tablet sales from the first decline in sales last quarter. The CAGR for total PC graphics from 2014 to 2017 is up to almost 3%. We expect the total shipments of graphics chips in 2017 to be 510 million units. In 2013, 454 million GPUs were shipped and the forecast for 2014 is 468 million.
The quarter in general
- AMD's shipments of desktop heterogeneous GPU/CPUs, i.e., APUs increased 10.5% from the previous quarter, and decreased 16% in notebooks. AMD's discrete desktop shipments decreased 19% and notebook discrete shipments increased 10%. The company's overall PC graphics shipments decreased 7%.
- Intel's desktop processor embedded graphics (EPGs) shipments decreased from last quarter by 0.3%, and notebooks increased by 18.6%. The company's overall PC graphics shipments increased 11.6%.
- Nvidia's desktop discrete shipments increased 24.3% from last quarter; and the company's notebook discrete shipments increased 3.5%. The company's overall PC graphics shipments increased 12.9%.
- Year-to-year this quarter AMD's overall PC shipments decreased 24%, Intel increased 19%, Nvidia decreased 4%, and the others essentially are too small to measure.
- Total discrete GPU (desktop and notebook) shipments from the last quarter increased 6.6%, and decreased 7.7% from last year. Sales of discrete GPUs fluctuate due to a variety of factors (timing, memory pricing, etc.), new product introductions, and the influence of integrated graphics. Overall, the trend for discrete GPUs has increased with a CAGR from 2014 to 2017 now of 3%.
- Ninety nine percent of Intel's non-server processors have graphics, and over 66% of AMD's non-server processors contain integrated graphics; AMD still ships integrated graphics chipsets (IGPs).
For PC and mobile device related companies small and large, new to the industry or established, it is critical to get a proper grip on this highly complex technology and understand its future direction. In this detailed 50-page data-based report, JPR provides all the data, analysis and insight needed to clearly understand where this technology is today and where it's headed. This fact and data-based report does not pull any punches: frankly, some of the analysis and insight may prove to be shocking.
Findings include discrete and integrated graphics (CPU and chipset) for Desktops, Notebooks (and Netbooks), and PC-based commercial (i.e., POS) and industrial/scientific and embedded. This report does not include the x86 game consoles, handhelds (i.e., mobile phones), x86 Servers or ARM-based Tablets (i.e. iPad and Android-based Tablets), or ARM-based Servers. It does include x86-based tablets, Chromebooks, and embedded systems.
58 Comments on Big Swing in Market Share From AMD to NVIDIA: JPR
B. Its not like they were not planning on making new cards, but unlike Nvidia just going with what currently is available unlike what they originally planned, AMD is sticking to the newest tech and are waiting for 20nm to be feasible.
Intel's iGPUs are very crude. The SIMD components are essentially stripped down x86 cores (a miniaturized Larrabee). These are not very space-efficient (≠ energy-efficient), and Intel can't build GPUs that compete with GM204 or Hawaii in these 3~5 years, not even having a fab-process advantage over TSMC.Yeah, they're busy making reeeally sure there's no crippling bug or design flaw with those chips, which could warrant a recall.
With the GTX970 and 980, nVidia has simply decided to strip practically all the rest of the double precision floating point circuitry out of the chip (which is like a car manufacturer pulling the back seats out of their new model and selling it as a 'feature' that the car is now lighter and faster) and release a new generation on the old 28nm process rather than waiting in line behind AMD at Taiwan Semiconductor for 20nm, because nVidia is always second in line after AMD, being as how ATI has a longer history with TSM. It's a clever move on nVidia's part, and one which nVidia supporters have been lapping up, but AMD's nextgen 390-series GPUs are on schedule for early 2015, and will almost certainly raise the bar past the GTX980, just like the 7970 did to the GTX580, and the 5870 did to the GTX480.
Before you have a crying party for poor AMD, remember that AMD has revenue streams beyond graphics cards, too, and it's only a matter of a year or two before Intel and AMD both crush nVidia's main add-in graphics business with APUs and poor nVidia with no x86 license. AMD's next gen CPU, K12, will also be out near the end of 2015. When AMD's APUs have stacked GDDR5 memory on die, with Radeon 300-series graphics and K12 cores, most people will not even have to bother with discrete graphics cards. Only the most fanatical nVidiots will continue to spill huge wads of cash on 2 and 3 ultra-uber cards, and the game makers won't even be thinking about them.
Don't feel bad for AMD, worry about nVidia, and what they're going to do when add-in graphics cards go the way of add-in sound cards and add-in ethernet cards. They'd better hope AMD doesn't start winning even more of their professional graphics card market at the same time.
Jim Keller is building this nextgen AMD CPU, and remember, he's the engineer who designed the AMD K8, otherwise known as the AMD Athlon 64, the last AMD CPU that clearly beat Intel's best at the time. If anybody can do it, he can.
To the point at hand, I don't care. AMD releasing first, getting a minor lead, and then losing the lead to Nvidea almost immediately isn't a good thing. Being the first to market doesn't mean you've got the best product, just that you pushed it out faster.
To the same ends, Zen is years away. I'll believe it can compete with Intel the second that it materializes. I was promised that Bulldozer would compete with Intel years ago, so consider me jaded when it comes to open promises. The second AMD can compete is the second they deserve respect. Right now Nvidea is eating their lunch when it comes to gaming, Intel is eating their lunch when it comes to CPUs, and ARM is eating their lunch when it comes mobile devices. If the APU didn't find a home in the tablet market, and if the consoles didn't rely on the APU as a cheap pre-designed processor, AMD would be in really hot water.
The assertion that Nvidea is competing by cutting out features is interesting, but a fallacy. If a GPU, primarily used for gaming, no longer has calculation ability who cares? The people who use a GPU for computation still have the previous generation of card, which compete well with AMD's most recent offering. Gamers will never know the difference, but the much lighter cost will allow them to buy into the next generation. While I'm not in favor of an upgrade which removes features, this seems like a win-win for the consumers. Lower prices, and performance improvements, mean good things for consumers.
I'm sure your response is going to be something about my biases. Good luck there. I've built AMD systems, Intel systems, and used both flavors of GPU. AMD has always delivered a slightly worse product, at a much better price. That's definitely still true in the CPU market, but the GPU market is another story. Sales figures show this conclusion is extremely valid, and arguing to the opposite point is a fool's errand. Whether you like Nvidea/Intel or not, they're offering a better product right now. AMD cannot compete well with this, and their sales demonstrate that fact. It doesn't matter how AMD will be competing in 5 years, if they can't survive the next 4 due to massively diminished sales.
new HBM technology would be very effective in GPU.
www.sisoftware.eu/rank2011d/show_run.php?q=c2ffcdfbddbcdde0d1e8d8ebd9e9cfbd80b096f396ab9bbdcef3c3&l=en
www.sisoftware.eu/rank2011d/top_run.php?q=c2ffcdfbddb0dde0d7f183be8ea8c1fccdeb83be8badd5e8d9ff9affc2f0c7f4ccea99a494&l=en
www.bitsandchips.it/9-hardware/4994-dall-isscc-nuove-informazioni-su-carrizo-hbm-in-arrivo
www.bitsandchips.it/9-hardware/4994-dall-isscc-nuove-informazioni-su-carrizo-hbm-in-arrivo
Carrizo APU as well.
Sound's like you'll be an unhappy camper when the 390X turns up on TSMC's 28nmHP node then
You really expect iGPs to consume the discrete GPU market? You obviously haven't been paying attention. Both vendors are pushing 4K (and above) gaming and higher graphical image quality settings. How many APU's are capable of playing games at 4K? how about path/ray tracing ? There's a reason that games outpace the ability of a single GPU to run them at max IQ/max resolution/max AA, and those reasons are Gaming Evolved and TWIMTBP. Hardware vendors aren't in the business of destroying their own markets.Yes, as if AMD have a record of sticking to CPU/APU schedules :rolleyes:Incorrect. Both ATI and Nvidia used TSMC's 250nm node in 1997-98, both have been continual customers of TSMC since (with small diversions to IBM, UMC, UICC, and Chartered). The reason Nvidia jumps onto a process node after ATI/AMD is because of getting burned in the past (TSMC's 130nm Hi-K process), and because once AMD subsumed ATI it had the ability to troubleshoot foundry process issues better thanks to AMD running its own fabs (at the time of course). I might also point out that Nvidia seems quick off the mark with 16nmFF+, so I wouldn't base a future case scenario solely from past history. As you yourself said, AMD of late have had a process lead (R700, Evergreen etc), but that lead has been eroding with every successive series.