A Timely Discovery: Examining Our AMD 2nd Gen Ryzen Results
by Ian Cutress & Ryan Smith on April 25, 2018 11:15 AM EST
Last week, we published our AMD 2nd Gen Ryzen Deep Dive, covering our testing and analysis of the latest generation of processors to come out from AMD. Highlights of the new products included better cache latencies, faster memory support, an increase in IPC, an overall performance gain over the first generation products, new power management methods for turbo frequencies, and very competitive pricing.
In our review, we had a change in some of the testing. The big differences in our testing for this review was two-fold: the jump from Windows 10 Pro RS2 to Windows 10 Pro RS3, and the inclusion of the Spectre and Meltdown patches to mitigate the potential security issues. These patches are still being rolled out by motherboard manufacturers, with the latest platforms being first in that queue. For our review, we tested the new processors with the latest OS updates and microcode updates, as well as re-testing the Intel Coffee Lake processors as well. Due to time restrictions, the older Ryzen 1000-series results were used.
Due to the tight deadline of our testing and results, we pushed both our CPU and gaming tests live without as much formal analysis as we typically like to do. All the parts were competitive, however it quickly became clear that some of our results were not aligned with those from other media. Initially we were under the impression that this was as a result of the Spectre and Meltdown (or Smeltdown) updates, as we were one of the few media outlets to go back and perform retesting under the new standard.
Nonetheless, we decided to take an extensive internal audit of our testing to ensure that our results were accurate and completely reproducible. Or, failing that, understanding why our results differed. No stone was left un-turned: hardware, software, firmware, tweaks, and code. As a result of that process we believe we have found the reason for our testing being so different from the results of others, and interestingly it opened a sizable can of worms we were not expecting.
An extract from our Power testing script
What our testing identified is that the source of the issue is actually down to timers. Windows uses timers for many things, such as synchronization or ensuring linearity, and there are sets of software relating to monitoring and overclocking that require the timer with the most granularity - specifically they often require the High Precision Event Timer (HPET). HPET is very important, especially when it comes to determining if 'one second' of PC time is the equivalent to 'one second' of real-world time - the way that Windows 8 and Windows 10 implements their timing strategy, compared to Windows 7, means that in rare circumstances the system time can be liable to clock shift over time. This is often highly dependent on how the motherboard manufacturer implements certain settings. HPET is a motherboard-level timer that, as the name implies, offers a very high level of timer precision beyond what other PC timers can provide, and can mitigate this issue. This timer has been shipping in PCs for over a decade, and under normal circumstances it should not be anything but a boon to Windows.
However, it sadly appears that reality diverges from theory – sometimes extensively so – and that our CPU benchmarks for the Ryzen 2000-series review were caught in the middle. Instead of being a benefit to testing, what our investigation found is that when HPET is forced as the sole system timer, it can sometimes a hindrance to system performance, particularly gaming performance. Worse, because HPET is implemented differently on different platforms, the actual impact of enabling it isn't even consistent across vendors. Meaning that the effects of using HPET can vary from system to system, as well as the implementation.
And that brings us to the state HPET, our Ryzen 2000-series review, and CPU benchmarking in general. As we'll cover in the next few pages, HPET plays a very necessary and often very beneficial role in system timer accuracy; a role important enough that it's not desirable to completely disable HPET – and indeed in many systems this isn't even possible – all the while certain classes of software such as overclocking & monitoring software may even require it. However for a few different reasons it can also be a drain on system performance, and as a result HPET shouldn't always be used. So let's dive into the subject of hardware timers, precision, Smeltdown, and how it all came together to make a perfect storm of volatility for our Ryzen 2000-series review.
Facebook
Twitter
RSS
242 Comments
View All Comments
ConstructionKing - Saturday, April 28, 2018 - link
Get ready to grab the best Flooring services in Ernakulam . We provide all the services of flooring including floor tile supply, professional flooring works etc at reasonable price. http://www.inframall.net/flooring.phpConstructionKing - Saturday, April 28, 2018 - link
Inframall undertakes all the roofing works including roofing sheet supply, roofing tiles supply, repair works and all other roofing related construction works at lowest rate.http://www.inframall.net/roofing.phpConstructionKing - Saturday, April 28, 2018 - link
Get experienced professionals from Kochi, Ernakulam and near locations to complete your plumbing requirements including man power and material supply. Click Details : http://www.inframall.net/plumbing.phpConstructionKing - Saturday, April 28, 2018 - link
Grab the best offers on electrical construction services. We carry professional electrical fitting services and fitting items at lower cost. Complete details @ http://www.inframall.net/electrical.phpConstructionKing - Saturday, April 28, 2018 - link
We Inframall do all the Repair And Maintenance Services for building and maintenance home constructions. Visit our website http://www.inframall.net/repair-maintenance.php to learn more.ConstructionKing - Saturday, April 28, 2018 - link
Find and choose the top best interior designers and architects for your home or commercial building constructions from our certified associate platform. We undertake all interior works at minimum rate in Ernakulam. http://www.inframall.net/interior-design.phpConstructionKing - Saturday, April 28, 2018 - link
Inframall in Ernakulam undertake all the painting works including material supply and man power. We aggregates all the professionals and material suppliers under single roof and makes the customers easy to completes their construction works.http://www.inframall.net/painting.phplyeoh - Saturday, April 28, 2018 - link
After so many decades and with so many transistors available can't Intel/AMD add better more efficient and accurate ways of getting time (e.g. monotonic time and also "real time")? They keep adding YetAnotherSIMD but how about stuff like this? For bonus points add efficient easy ways to set (and cancel) interrupts that will trigger after certain times.patrickjp93 - Monday, April 30, 2018 - link
It's a total misconception that any circuit can get these times more accurately and efficiently at the same time. There's no demand for it on a broad scale either.Bensam123 - Saturday, April 28, 2018 - link
This has been something the gaming community has known about and discussed for years, you can find posts about HPET all over popular forums. However, they aren't backed by any sort of meaningful data and much like with core parking and vpns for gaming, legitimate hardware testing websites generally have either turned up their nose at it or disregarded them as complete snake oil.I'm glad HPET is finally getting looked at in real depth. What was discussed here was just their real effect on quantifiable metrics, such as benchmarks, but what gamers discuss is their impact on stuttering or microstutters as well as hit registration in netcode (which is extremely dependent on timing). That wasn't looked at at all here. While this discussion was mainly focused on the difference between results between other websites and Intel vs AMD, I think that's not quite the right way to approach this. Rather it should be looked at the effect of timers in general on gaming as a whole.
The statement that you guys received from Intel pretty much makes it blatantly clear that no one really had any idea what was going on with the timers over there or that they really had a big impact on anything outside of synthetic results. Microsoft has just put band aids on top of band aids to keep everything running and it got to the point where it's no longer transparent to people who are buying hardware, people who are making hardware, or people who are developing software (beyond a few very niche groups) how they all interlock and intermingle with each other. I didn't until I did some digging and required even more to learn there were more timers besides HPET between multiple and sometimes vaguely related forum posts.
A higher resolution timer should be good, especially for video games, but the impact it has on the system because of it's crude and backwards implementation has made it such that it's basically just a synthetic cog that can't be used in practice. It makes you wish there was a solution that just put everything on the same page, hardware and software. I'm sure game developers who just lease a engine and then essentially make a mod have no idea what's going on here and developers who actually make the engine (Dice, Epic, Crytek) may not even know there is a problem in the first place.
I do hope Anand takes this a bit further with frame time benchmarks and maybe FCAT designed to look specifically at this. As was mentioned in this article, implementations even seem different across different motherboards, which is a very, very bad thing and should also be looked at. There is a lot of room for future articles here focused around this specific issue until there is some remote amount of standardization.
If you're looking for more interesting things to test - almost no one tests net code in video games, with the handful of people who do making arbitrary comparisons and really having no tools or benchmarks to work with, even though video games (especially highly competitive ones) are extremely dependent on such things, especially when you get into the top 10% of the player base. People just assume gaming code 'works' and that magical part of games are all created equal when it couldn't be any further from the truth. Net code is literally trying to hold together what is essentially a train wreck while trying to mask it from it's users as best as possible. Some games do it a lot better then others.