NotebookCHECK - Notebook Forum

English => News => Topic started by: Redaktion on April 10, 2021, 14:11:24

Title: AMD's Ryzen 9 5900 spotted on UserBenchmark, only ~5% slower than the more power-hungry 5900X
Post by: Redaktion on April 10, 2021, 14:11:24
Even though the non-X processors come with a 61.5% lower TDP, 700 Hz lower base clock and 100 Hz lower boost clocks, the performance loss is barely noticeable with the UserBenchmark suite versus the 5900X models. The Ryzen 9 5900 is an OEM-only model and is now available in select pre-built gaming rigs like the Alienware Aurora.


https://www.notebookcheck.net/AMD-s-Ryzen-9-5900-spotted-on-UserBenchmark-only-5-slower-than-the-more-power-hungry-5900X.531442.0.html
Title: Re: AMD's Ryzen 9 5900 spotted on UserBenchmark, only ~5% slower than the more power-hungry 5900X
Post by: Matthew Bogle on April 10, 2021, 18:44:02
TDP != power consumption, especially for AMD desktops. Don't pretend for a second that the stock power limit of the non-X is any different from the X.
Title: Re: AMD's Ryzen 9 5900 spotted on UserBenchmark, only ~5% slower than the more power-hungry 5900X
Post by: _MT_ on April 10, 2021, 19:52:38
QuoteWe were expecting to see a greater gap for single-core loads, due to the non-X model having a 700 Hz lower base clock, but this does not seem to influence scores that much, at least for the UserBenchmark suite. Boost clock is only 100 Hz lower, so that pretty much guarantees the multi-core performance is essentially the same on both models.
Really? Then you seem rather clueless as to how things work. If a power limit is going to be a problem, then it will be in an all-out multi-threaded load. Lower TDP does imply it might have a lower power limit, but not necessarily. Single-threaded loads typically can't hit limits on desktop-class processors. The limit is just too high for a single core to pull that much power. And whether base frequency matters vs. boost frequency depends on how long is the test and how exactly the boosting algorithm works. Time, temperature, power. If you have an algorithm primarily based on temperature and the same power limits, cooler will dictate performance. Irrespective of TDP. That's why TDP can be a self-fulfilling prophecy. Lower TDP will prompt you to buy a weaker cooler which will then result in the expected weaker performance.
Title: Re: AMD's Ryzen 9 5900 spotted on UserBenchmark, only ~5% slower than the more power-hungry 5900X
Post by: _MT_ on April 10, 2021, 20:16:51
Quote from: Matthew Bogle on April 10, 2021, 18:44:02
TDP != power consumption, especially for AMD desktops. Don't pretend for a second that the stock power limit of the non-X is any different from the X.
It seems like the biggest difference might be binning. Hence the lower boost.
Title: Re: AMD's Ryzen 9 5900 spotted on UserBenchmark, only ~5% slower than the more power-hungry 5900X
Post by: Logoffon on April 11, 2021, 05:18:30
If you're using Hz, then it must be 700,000,000 Hz lower base clock than 5900X and not 700 Hz, or it would be 3.699999300 GHz.

1000 Hz = 1 KHz
1000 KHz = 1 MHz
1000 MHz = 1 GHz
Title: Re: AMD's Ryzen 9 5900 spotted on UserBenchmark, only ~5% slower than the more power-hungry 5900X
Post by: _MT_ on April 11, 2021, 07:52:36
Quote from: Logoffon on April 11, 2021, 05:18:30
1000 Hz = 1 KHz
1000 KHz = 1 MHz
1000 MHz = 1 GHz
It's kHz. SI is case sensitive. And kilo- is k. K is kelvin. What you wrote is kelvin-hertz. And if you are wondering, hertz is a unit while Hertz is a family name.
Title: Re: AMD's Ryzen 9 5900 spotted on UserBenchmark, only ~5% slower than the more power-hungry 5900X
Post by: T-T on April 11, 2021, 22:07:52
Please stop looking to userbenchmark. It has been shown to be heavily biased in favor of intel. Multiple Youtubers and both intel and amd run reddit have banned links to it.