Stories based on Passmark data about large and rapid market share gains for AMD's Ryzen x86 processors are ill informed and raise expectations above reason.
Right before the holiday break in the U.S., stories trumpeted significant market share gains for AMD’s Ryzen processors. At first glance, these stories made sense. AMD launched Ryzen earlier this year, its first major new architecture in nearly a decade, to mostly-positive reviews. If ever there was a product more hyped and interesting for the consumer and DIY markets, I’d be hard pressed to name it.
The source of this data was Passmark, a benchmark that is free, simple to run, and provides a no-nonsense, score-based result that is easy to interpret. The scores submitted using Passmark are publicly viewable and the company often combs the data to find interesting trends when it can. This week the data was updated, adding information about a shift in the market share between Intel and AMD.
One story said the Passmark data “suggests Ryzen is chipping away at Intel's considerable lead with the enthusiasts.” Another stated, “AMD has managed to claim back almost 10 percent of the market from Intel,” and another goes as far as telling readers that “AMD has gained 5.2%, moving up from a market share of 20.6% the previous quarter, to reach 25.8%.”
Data like this from Passmark can be incorrectly interpreted as sales or market share changes. (Image: Passmark)
There is no doubt in my mind that with the release of Ryzen AMD is gaining market share and will continue to do so for the foreseeable future. But some of the conclusions based on Passmark data are simply ill informed and serve only to raise expectations above reason for AMD ahead of its second quarter financial report.
Passmark benchmark submissions count the number of times scores are generated with a PC and processor. There is no limit to the number of scores that can be submitted from any one system, CPU or individual. The scores are not monitored by anyone from Passmark and are not culled for duplicate entries--that’s not the purpose of the system.
What the results actually tell us is that OEMs, gamers, reviewers, enthusiasts, and others are all very excited about their new Ryzen hardware and are running benchmarks on new system builds. As the community and individuals learn about the intricacies of the Ryzen platform, including memory speeds, etc., the tests get run, and re-run.
On the other hand, Intel processors are known quantities. They have been in the market and in enough hands that Passmark scores are stable and reliable, and there is no reason to run them again. So the Passmark Ryzen data reflects the increase in submission of Passmark results, it is not an indicator of market share.
The PC market buys roughly 75 million units per quarter. For AMD to capture 10% of that in a single quarter, it would need to ship and sell nearly eight million more units from April through June than it did in January through March. Correspondingly, Intel would be shipping that many fewer processors.
If Intel realized that it suddenly lost eight million processor sales at a rough average selling price of $150, it would be reporting to shareholders a sudden billion-plus dollar revenue drop. Last quarter Intel’s Client Computing Group reported $8 billion in revenue. The drop some say Passmark indicates would be equivalent to more than a 12.5% decline.
Benchmark submissions from Passmark are a positive data point for AMD. More benchmarks being run is a concrete indicator of improved sales. But the assumptions being made by many in the media, and spread to the sometimes-gullible financial market, are dramatically exaggerated. This narrative is bad for both Intel and AMD, so it is best to try to get the story straight.
--Ryan Shrout is the founder and lead analyst at Shrout Research and the owner of PC Perspective.