by Bob Cramblitt
There’s a line of reasoning that goes like this: If it looks like a duck, swims like a duck, and quacks like a duck, then it’s probably a duck. The assumption is that something can be identified by its characteristics and unique behavior.
But, as we all know, looks and behavior can be deceiving.
Such is the case with a PC versus a workstation. They can look the same, often process the same, display the same, but they are two different beasts.
Defining a workstation
Jon Peddie Research (JPR) identifies the following characteristics of a workstation:
- Must have a workstation-class CPU such as an AMD Ryzen-WS or Intel Xeon
- Must have a workstation-class graphics card such as an AMD Radeon Pro or Nvidia Quadro
- Must have error-correcting code (ECC) memory
- Must have application-specific, certified drivers
- Must have Windows Pro or an equivalent Linux operating system
- Should have a minimum of 32 GB RAM
- Must be ultra-reliable
Lack of clarity persists
Unfortunately, despite dozens of articles and white papers documenting the differences between a PC and a workstation, the perception that they are interchangeable still exists.
“Part of the reason for the lack of clarity on what is needed to qualify as a workstation is due to the willingness of some brands to label PCs as workstations,” says Kathleen Maher, JPR vice president.
A recent survey by JPR of hundreds of both workstation and non-workstation users shows how this lack of clarity persists, despite efforts of system and component vendors to delineate the differences. Some key findings:
- Nearly 19% of respondents don’t know the difference between a gaming and a workstation graphics card or say that it doesn’t matter.
- More than 21% of respondents either don’t know the difference between a consumer-level and a workstation-level CPU or say it doesn’t matter. Of the 69% who have an opinion, only 32% felt a workstation needed what is considered a workstation-class CPU such as an AMD Ryzen or Intel Xeon.
- Respondents largely discount the importance of ISV certification, considered an essential element of a professional workstation. Only about half of those surveyed think that certification is critical or even good to have, while a surprising 30% say it isn’t needed.
- Fifty-five (55) percent of respondents do not identify ECC as an essential differentiator between a PC and a professional workstation.
- Extra reliability designed into workstations and the rigorous testing during the R&D process does make a difference to workstation users, but the majority of non-workstation users think a commercial-grade system is good enough to be considered a workstation.
Performance pros weigh in
An informal group of professionals involved in performance benchmarking find the same types of misconceptions as those indicated by the JPR survey.
“Without question, the most common misconception I see is that workstation graphics cards are the same as gaming graphics cards, except that they cost more,” says Rob Williams, editor-in-chief of Techgage. “There’s a lot more to it than that. There are perks of workstation GPUs that are less obvious, like blanket 10-bit OpenGL support, ECC, more memory, and of course, some optimizations for professional workloads in applications such as Catia and Solidworks.”
Trey Morton, chair of the SPEC Application Performance Characterization (SPECapc) subcommittee, agrees with Williams’ assessment.
“There are a lot of professionals who want to use a high-end gaming card for workstation applications,” he says. “While the performance is definitely there for games, the consumer-level graphics cards don’t support the full functionality of workstation applications or go through the ISV certification process that the professional cards require, thus support from the ISV could be limited to non-existent.”
Joel Hruska, senior editor of ExtremeTech, concurs that certification and support are critical differences.
“The difference between workstation and consumer hardware, in my view, boils down to software support and certification. It’s having a driver that’s been specifically validated to work with a product, such that if you do encounter a problem, you’ll be able to actually speak to someone at the company as opposed to being shoved into the black hell known as Tier 1 Online Chat-Based Product Support.”
Kevin O’Brien, lab director for StorageReview.com, says that misconceptions about PC versus workstation requirements extend to CPUs.
“Workstation-class systems generally include highly reliable server processors, which can vary from low clock speed and high core count to high clock speed and lower core count. Many consumer applications still seem to rely on single-threaded performance in certain areas, which can respond differently depending on the processor selected.”
Allen Jensen, vice chair of the SPEC Graphics and Workstation Performance Group (SPEC/GWPG) and the SPECapc subcommittee, notes that the higher price of certified graphics cards is justified by testing, support, and ongoing updating of drivers.
“Beyond certification, graphics card vendors subject workstation-class products to tens-of-thousands of test cases, thousands of QA days, and hundreds of hardware configurations and applications before they make it to a professional’s desktop. A card that has not been thoroughly tested and certified can cause a crash, resulting in lost work and a search for answers. ”
What’s your risk tolerance?
A tolerance for risk balanced against the critical nature of the work seems to be a key indicator of whether a company invests in a workstation versus a gaming PC. Some of the questions that decision-makers relying on high-end professional applications such as Catia, Solidworks, NX, Creo, or Maya should ask include:
Can your company afford to risk problems associated with an uncertified graphics card or lack of ECC memory that could derail a project for days or weeks?
Can your company afford to have a highly paid engineer sitting idly by while an issue is sorted out without vendor support?
Do the savings from purchasing a PC versus a workstation really represent a return-on-investment when all factors are considered?
Based on the JPR survey, the answers to the questions above seem to be a tacit yes for many professional computer users. Or it could be a case of discounting a potential problem until it hits home in the user’s workplace.
Regardless of the reasoning of the decision-makers, who purchase PCs for workstation-level applications, there seems to be unanimous agreement among analysts, hardware reviewers, and performance professionals: There is clearly a difference and not recognizing that fact could lead to dire consequences.
Bob Cramblitt is communications director for SPEC. He writes frequently about performance issues and digital design, engineering and manufacturing technologies.
To find out more about graphics and workstation performance benchmarking and the availability of free benchmarks, visit the SPEC/GWPG website, subscribe to the SPEC/GWPG enewsletter, or join the Graphics and Workstation Benchmarking LinkedIn group: https://www.linkedin.com/groups/8534330.