Ticked Off: Clock Bug Leads to False Ryzen Benchmark Scores

[Admin]

A persistent bug in Windows 10 means that 3rd Gen Ryzen chips could give inaccurately high or low test scores.

Ticked Off: Clock Bug Leads to False Ryzen Benchmark Scores : Read more

 

[Kiril1512]

Very interesting article! Thank you Tom's Hardware for this.

 

[paul prochnow]

Will there be a settled upon acceptable BenchMark? So GeekBench is NFG too?😵

 

[deltakprime]

Thank you for this comprehensive and most informative article, and thanks to Matthias Zronek.

In typical fashion of this article, you have been a reputable and reliable standard for so long now that I have to stop and think of when I first consulted Tom's Hardware... probably the early '90's?

In any case, you remain indisputably an invaluable resource, and I want to openly salute everyone at this fine organization for their outstanding performance.

Keep it up... you're the best!

 

[Diceman_2037]

Writer is incorrect, HPET needs only be enabled in the bios - it is not required that the OS be forced to use HPET.

Any writer that says bcdedit /set useplatformclock yes is needed to resolve this should be entirely ignored, they have clearly failed to do their research.

 

[jimmysmitty]

Writer is incorrect, HPET needs only be enabled in the bios - it is not required that the OS be forced to use HPET.

Any writer that says bcdedit /set useplatformclock yes is needed to resolve this should be entirely ignored, they have clearly failed to do their research.

The article is neither wrong nor 100%. Its actually both. You have to verify the BIOS supports it and make sure its enabled then you have to force Windows to use it:

Try changing HPET settings to improve your PC's performance - gHacks Tech News

Find out how to detect timer settings on a Windows PC, to change them to improve performance, and to measure the performance using benchmarks.

www.ghacks.net

 

[Darksurf]

So basically this is a Windows bug meaning all Linux benchmarks such as from phoronix show true results (just without optimization coming in later kernels).

 

[Arbie]

Clock Bug Leads to False Ryzen Benchmark Scores

Actual fact: If you've adjusted the base clock of your Ryzen system, benchmark scores will be affected.

It's hard to avoid the perception that your article title is clickbait, and anti-AMD clickbait at that. At most it should have said "can lead to false Ryzen benchmark scores".

As it is, you've successfully given the impression that all Ryzen benchmark scores are suspect. Intel marketing is surely happy to see that in a headline.

Does all this somehow not even occur to you? Really?

 

[jimmysmitty]

Clock Bug Leads to False Ryzen Benchmark Scores

Actual fact: If you've adjusted the base clock of your Ryzen system, benchmark scores will be affected.

It's hard to avoid the perception that your article title is clickbait, and anti-AMD clickbait at that. At most it should have said "can lead to false Ryzen benchmark scores".

As it is, you've successfully given the impression that all Ryzen benchmark scores are suspect. Intel marketing is surely happy to see that in a headline.

Does all this somehow not even occur to you? Really?

In that context the word "can" would indicate the possibility of it occurring when the fact is that it will occur due to the bug listed in this article.

I think the title works perfectly fine. The fact is that this specific bug means benchmark scores will be affected under certain conditions which then get explained in said article.

 

[dansthunder]

I would agree -- Seems like this article was written by the Intel Marketing Team. Windows 10 bug, affects both Intel & AMD, except maybe the later generation Intel, doesn't say how or using what magic. A better title would be Windows 10 dynamic tic prevents accurate benchmarks. Period. Nah, let's turn it into click bait.

 

[Arbie]

... under certain conditions which then get explained in said article.

My point exactly. If you read the article, you get the facts. If you read the headline, you get misled. Ergo, most people scanning headlines on Tom's will be misled.

 

[jimmysmitty]

My point exactly. If you read the article, you get the facts. If you read the headline, you get misled. Ergo, most people scanning headlines on Tom's will be misled.

Headlines should grab attention. Enough so that a person should open the article and read it. If they don't then there isn't much they can do.

 

[NightHawkRMX]

Microsoft needs to step it up.

Bug after bug. If your paying over $100 for an operating systrm, you would expect it to work better than a free linux based os.

Every day its something new pertaining to ryzen.
Windows scheduler issues, Windows power management issues, high idle cpu usage issues.

I cant see how some of these are hard to fix.

When Windows detects a ryzen cpu it should change some things. Adapt the sceduler and power settings for good performance but also efficiency

 

[TJ Hooker]

Microsoft needs to step it up.

Bug after bug. If your paying over $100 for an operating systrm, you would expect it to work better than a free linux based os.

Every day its something new pertaining to ryzen.
Windows scheduler issues, Windows power management issues, high idle cpu usage issues.

I cant see how some of these are hard to fix.

When Windows detects a ryzen cpu it should change some things. Adapt the sceduler and power settings for good performance but also efficiency

This 'bug' is a complete non-issue for users. It only manifests if you change your BCLK within Windows, which isn't something there's really any reason to do. And even then, the impacts of the bug are non-existent unless you're doing something that relies on precise CPU timing.

 

[NightHawkRMX]

That doesnt speak for all of the other issues.

 

[TJ Hooker]

That doesnt speak for all of the other issues.

Sure, but those other issues have nothing to do with this article.

 

[x3roxide]

OK, so I have a stock windows 10 pro, x64 system
Ryzen 3700x and I downloaded and installed all motherboard and chipset drivers.
The only "overclocking" I did was load up my xmp profiles on the ram and tweak it a little to get a high infinity fabric ratio 1:1

downloaded and installed benchmate 0.9.3 (big) and did some benchmarks.
It passed all timings without errors, both in BenchMate and cpu-z timer test.

my results...
cinebench R15 score: 2167
cinebench R20 score: 4899

I've had this sort of error with Intel in the past. What I've come to the conclusion is that motherboard manufacturers will tinker ever so slightly with this timing to get an edge in benchmarks over their rivals. key point being motherboard manufacturers... not cpu!

 

[TJ Hooker]

OK, so I have a stock windows 10 pro, x64 system
Ryzen 3700x and I downloaded and installed all motherboard and chipset drivers.
The only "overclocking" I did was load up my xmp profiles on the ram and tweak it a little to get a high infinity fabric ratio 1:1

downloaded and installed benchmate 0.9.3 (big) and did some benchmarks.
It passed all timings without errors, both in BenchMate and cpu-z timer test.

my results...
cinebench R15 score: 2167
cinebench R20 score: 4899

I've had this sort of error with Intel in the past. What I've come to the conclusion is that motherboard manufacturers will tinker ever so slightly with this timing to get an edge in benchmarks over their rivals. key point being motherboard manufacturers... not cpu!

Err, what is the "error" you're referring to? Both those scores are right in line with what you'd expect from a 3700X.

Edit: And your 2nd link is broken.

 

[jimmysmitty]

Microsoft needs to step it up.

Bug after bug. If your paying over $100 for an operating systrm, you would expect it to work better than a free linux based os.

Every day its something new pertaining to ryzen.
Windows scheduler issues, Windows power management issues, high idle cpu usage issues.

I cant see how some of these are hard to fix.

When Windows detects a ryzen cpu it should change some things. Adapt the sceduler and power settings for good performance but also efficiency

Scheduler issues are up to both Microsoft and AMD to fix. AMD should be working with Microsoft to optimize the scheduler for their CPUs before launch. This happened with their FX series of CPUs where Microsoft had to retool it to handle the FX modules differently since loading up one module then the next resulted in worse performance rather than loading one pipeline per module then the second one in each.

OK, so I have a stock windows 10 pro, x64 system
Ryzen 3700x and I downloaded and installed all motherboard and chipset drivers.
The only "overclocking" I did was load up my xmp profiles on the ram and tweak it a little to get a high infinity fabric ratio 1:1

downloaded and installed benchmate 0.9.3 (big) and did some benchmarks.
It passed all timings without errors, both in BenchMate and cpu-z timer test.

my results...
cinebench R15 score: 2167
cinebench R20 score: 4899

I've had this sort of error with Intel in the past. What I've come to the conclusion is that motherboard manufacturers will tinker ever so slightly with this timing to get an edge in benchmarks over their rivals. key point being motherboard manufacturers... not cpu!

Motherboard manufatures have always had slight differences. Some will have slightly higher BCLK than others while some may be just a hair under the stock 100MHz.

However this bug itself is specific to CPUs and not the boards themselevs. Most board, unles you load some software, will not be changing the BCLK in Windows.

 

[x3roxide]

Scheduler issues are up to both Microsoft and AMD to fix. AMD should be working with Microsoft to optimize the scheduler for their CPUs before launch. This happened with their FX series of CPUs where Microsoft had to retool it to handle the FX modules differently since loading up one module then the next resulted in worse performance rather than loading one pipeline per module then the second one in each.



Motherboard manufatures have always had slight differences. Some will have slightly higher BCLK than others while some may be just a hair under the stock 100MHz.

However this bug itself is specific to CPUs and not the boards themselevs. Most board, unles you load some software, will not be changing the BCLK in Windows.

so why was I able to pass at stock settings?
the BCLK was not altered in any way, shape or form on my system.

edit
also, before thinking I'm just blindly loyal. I have been very vocal about the boost not actually reaching it's advertised speeds... and have had to argue with people saying "who cares, it's only 50mhz"

only recently AMD have admitted a bug in their current bios not allowing for max boost.
I call a spade a spade and honestly, I have not found an issue with the BCLK at all.

BUT, I am on the latest Win10 1903 x64 with all drivers up-to-date.

 

[jimmysmitty]

so why was I able to pass at stock settings?
the BCLK was not altered in any way, shape or form on my system.

You answered your own question. It was stock settings. This bug is present if you change the BCLK in Windows manually. Say the board has a set BCLK of 100 and you boot into Windows and change the BCLK to 105 then run the benchmarks it could perform 5% better than if not changed.

 

[x3roxide]

You answered your own question. It was stock settings. This bug is present if you change the BCLK in Windows manually. Say the board has a set BCLK of 100 and you boot into Windows and change the BCLK to 105 then run the benchmarks it could perform 5% better than if not changed.

I re-read the article to see what I missed

actually it would be 5% worse

“Let’s say you boot your i7-4770K with a bclock (base clock) of 104 MHz into Windows 10, reduce it to 98 MHz with your favorite OC tool and run a benchmark. The measured time for the run would be 6% shorter. It also works in the other direction: Booting with 98 MHz and overclocking to 104 MHz would stretch the time by 6% ending up in a result that’s possibly worse than without overclocking.” Zronek explained.

Therefore
what you actually need to do is increase your bclk in bios to say 104% (which is dangerous btw - hdd failure/corruption) and then use a tool in windows to lower the bclk to say 98% for a net benefit of 6%

sounds more like an exploit than a bug to me. Important if you're going for world benchmark scores, but then again you'd also be on LN2.

For everyone else, modifying the bclk is not worth the hassle,

btw to confirm, I raised my bclk to 101mhz and still passed, because the clock has not changed from boot to runtime. I will not go higher because I don't want my hdd to corrupt.

 

[gdesigndog77]

I agree with jimmy smitty and Danthunder. This thing couldn't be written more "click-baitish " in favor of Intel processors than if it was written by the marketing people at Intel. Besides it sounds like thew whole thing was initiated by Intel/Microsoft effort when Windows 8 was introduced. Why does the headline read like all AMD Ryzen processors have questionable timing outcomes. Wait let me guess. Has AMD been getting some deservedly good press too long. Although the article thoroughly covered the topic. Shame on Tom's Hardware for running the article with that headline!!

 

[gdesigndog77]

I would agree -- Seems like this article was written by the Intel Marketing Team. Windows 10 bug, affects both Intel & AMD, except maybe the later generation Intel, doesn't say how or using what magic. A better title would be Windows 10 dynamic tic prevents accurate benchmarks. Period. Nah, let's turn it into click bait.

I agree with jimmy smitty and Danthunder. This thing couldn't be written more "click-baitish " in favor of Intel processors than if it was written by the marketing people at Intel. Besides it sounds like thew whole thing was initiated by Intel/Microsoft effort when Windows 8 was introduced. Why does the headline read like all AMD Ryzen processors have questionable timing outcomes. Wait let me guess. Has AMD been getting some deservedly good press too long. Although the article thoroughly covered the topic. Shame on Tom's Hardware for running the article with that headline!!

 

[jimmysmitty]

I agree with jimmy smitty and Danthunder. This thing couldn't be written more "click-baitish " in favor of Intel processors than if it was written by the marketing people at Intel. Besides it sounds like thew whole thing was initiated by Intel/Microsoft effort when Windows 8 was introduced. Why does the headline read like all AMD Ryzen processors have questionable timing outcomes. Wait let me guess. Has AMD been getting some deservedly good press too long. Although the article thoroughly covered the topic. Shame on Tom's Hardware for running the article with that headline!!

Its been an issue for Intel and AMD since Windows 8 however it still affects newer AMD CPUs and HWBot is the one who banned results from AMD Ryzen CPUs. If there is a software or firmware fix applied I am sure it will be reversed.