AMD Ryzen is Smoother Misconception Benchmark & Explanation (Ryzen vs. Intel for Gaming)
video description
Shame how intel in the long run decided to use a mountain of speed hacks with little regard for how they would affect security, now the chickens are coming home to roost.
Date: 2020-08-13
Related videos
Comments and reviews: 9
Hickeroar
Bottom line: If you're building a dedicated gaming PC, dollar for dollar, Intel still wins in most cases. 10600 vs 3700 makes this obvious. Roughly the same price, but the intel averages about 10% higher framerates. If you're doing productivity tasks that require heavy multi-core processing loads, the scenario is reversed, and probably a lot more than 10% in AMDs favor.
Building a dedicated gaming rig with AMD just doesn't make a lot of sense to me. Even if you pay more, the investment per dollar yields much higher dividends at the upper end of the pricing curve when you go with intel.
Something to think about: People are willing to pay nearly double the price for a 2080Ti vs a 2080Super, and that gets them 10% better performance, give or take. That's a HORRIBLE value for dollar ratio, and people do it all the time. But the AMD fans only want to talk about value when it comes to AMD. Sure you might save 30- 50 in some cases, but you're losing a full 10% of the gaming performance in many cases. They probably already made a much worse value decision just choosing a video card, and somehow they were totally fine with it there.
You can easily value your way into losing a combined 30% (or more) of your performance over spending a couple hundred more.
reply
Bottom line: If you're building a dedicated gaming PC, dollar for dollar, Intel still wins in most cases. 10600 vs 3700 makes this obvious. Roughly the same price, but the intel averages about 10% higher framerates. If you're doing productivity tasks that require heavy multi-core processing loads, the scenario is reversed, and probably a lot more than 10% in AMDs favor.
Building a dedicated gaming rig with AMD just doesn't make a lot of sense to me. Even if you pay more, the investment per dollar yields much higher dividends at the upper end of the pricing curve when you go with intel.
Something to think about: People are willing to pay nearly double the price for a 2080Ti vs a 2080Super, and that gets them 10% better performance, give or take. That's a HORRIBLE value for dollar ratio, and people do it all the time. But the AMD fans only want to talk about value when it comes to AMD. Sure you might save 30- 50 in some cases, but you're losing a full 10% of the gaming performance in many cases. They probably already made a much worse value decision just choosing a video card, and somehow they were totally fine with it there.
You can easily value your way into losing a combined 30% (or more) of your performance over spending a couple hundred more.
reply
LunarVVolf
I love your videos as of late being so biased you're really looking for topics to take the heat off Intel's recent pitfalls. For a minority that is AMD users you generalize so much in paint everyone with the same brush for simply having made a hardware choice. Am I a complete AMD nutjob no there's a 1080TI in my case and happy with it to skip the RTX 2k series altogether. Just looking at your videos in the last 6 months you trash on AMD more frequently and explicitly dedicating entire episodes to the critique while you treat Intel with kid gloves and point out Intel's flaws in short segments within videos covering a broad set of topics. Intel is still the larger company and still has the track record for unethical practices ranging from screwing its employee's retirement 401k plans with risky investing, anti-trust violations in each decade of its existence, using an affiliate company to write favorable reviews whilst using skewed test parameters, using an industrial chiller at a press demo, the security exploits and malware seemingly every month. Also who was it that sat on their ass and refused to innovate it wasn't AMD or its users keeping them from doing so. Its Intel's fault for being the position they're in. Let that sink in.
reply
I love your videos as of late being so biased you're really looking for topics to take the heat off Intel's recent pitfalls. For a minority that is AMD users you generalize so much in paint everyone with the same brush for simply having made a hardware choice. Am I a complete AMD nutjob no there's a 1080TI in my case and happy with it to skip the RTX 2k series altogether. Just looking at your videos in the last 6 months you trash on AMD more frequently and explicitly dedicating entire episodes to the critique while you treat Intel with kid gloves and point out Intel's flaws in short segments within videos covering a broad set of topics. Intel is still the larger company and still has the track record for unethical practices ranging from screwing its employee's retirement 401k plans with risky investing, anti-trust violations in each decade of its existence, using an affiliate company to write favorable reviews whilst using skewed test parameters, using an industrial chiller at a press demo, the security exploits and malware seemingly every month. Also who was it that sat on their ass and refused to innovate it wasn't AMD or its users keeping them from doing so. Its Intel's fault for being the position they're in. Let that sink in.
reply
wolfajk
I've not watched the entire video yet, but I'm a [backend] developer and here's a developers perspective: your machine will run processes, processes are a collection of memory and a thread schedule, this is why you might see 30 threads taken up by a program. They're not taken up, it just uses 30 threads in it's cycle, 20 of them are probably the same thread. With that said, how many threads becomes less important now, so does speed? Well threads must run parallel or synchronous, so, if not all cores are at the same speed, then not all threads can run parallel. This means we're forced to be synchronous and that just 'feels' slower for a user. Now where games are concerned, their thread schedules are are predominantly event based. This means they'll only use 2-4 actual threads, and have an event queue. In this case speed is heavily dependent on how it'll perform. But typically a normal program will utilise what the developer thinks they need. These days we aren't as worried about a machine running out of memory for a given process, so we'll technically follow bad practices and just spawn threads all over the place. In instances like this, I can see how more threads would perform better. Hope this helps anyone just curious.
reply
I've not watched the entire video yet, but I'm a [backend] developer and here's a developers perspective: your machine will run processes, processes are a collection of memory and a thread schedule, this is why you might see 30 threads taken up by a program. They're not taken up, it just uses 30 threads in it's cycle, 20 of them are probably the same thread. With that said, how many threads becomes less important now, so does speed? Well threads must run parallel or synchronous, so, if not all cores are at the same speed, then not all threads can run parallel. This means we're forced to be synchronous and that just 'feels' slower for a user. Now where games are concerned, their thread schedules are are predominantly event based. This means they'll only use 2-4 actual threads, and have an event queue. In this case speed is heavily dependent on how it'll perform. But typically a normal program will utilise what the developer thinks they need. These days we aren't as worried about a machine running out of memory for a given process, so we'll technically follow bad practices and just spawn threads all over the place. In instances like this, I can see how more threads would perform better. Hope this helps anyone just curious.
reply
Samuel
To be fair, I think a lot of people with stuttery inconsistent Intel systems may just be setting their uncore ratio too high.
The culture of higher number is better, and synthetic benchmarking numbers lead me to run a 9900KS at 5.2GHz with an uncore of 4.9Ghz. Whilst this could pass 8hrs of OCCT/Realbench without issue, and the synthetic benchmarks looked great, running the uncore this high introduced stutter and frame inconsistency/frame drops whilst gaming. A problem which was incrementally improved as I knocked the uncore down, until it was completely resolved at its stock setting (4.3GHz).
I have also observed this same behaviour on a 9900KF and a 7700K, however I appreciate this is still a very small sample size.
If you're on an Intel CPU and experiencing stuttering/inconsistency issues, give it a go; see if it helps!
I haven't been able to observe any issues with running a higher uncore outside of gaming (although gaming and benchmarking are admittedly my only two use cases).
reply
To be fair, I think a lot of people with stuttery inconsistent Intel systems may just be setting their uncore ratio too high.
The culture of higher number is better, and synthetic benchmarking numbers lead me to run a 9900KS at 5.2GHz with an uncore of 4.9Ghz. Whilst this could pass 8hrs of OCCT/Realbench without issue, and the synthetic benchmarks looked great, running the uncore this high introduced stutter and frame inconsistency/frame drops whilst gaming. A problem which was incrementally improved as I knocked the uncore down, until it was completely resolved at its stock setting (4.3GHz).
I have also observed this same behaviour on a 9900KF and a 7700K, however I appreciate this is still a very small sample size.
If you're on an Intel CPU and experiencing stuttering/inconsistency issues, give it a go; see if it helps!
I haven't been able to observe any issues with running a higher uncore outside of gaming (although gaming and benchmarking are admittedly my only two use cases).
reply
blake112x
I ended up buying a Ryzen 3700x over Intel at the time.
1) I was upgrading from i5 6600k (stock) and noticed with games like RDR2 - it did not respect 4 cores 4/4 and was having a unsmooth experience with some games.
2) Intel 9th Gen series was only available at the time and they dropped hypertherding from there i5/i7 series. And reviews seems to be suggesting that they where not as smooth becuase the lack of hypertherding.
So I took a chance on a AMD build and it has been okay, no real complaints. Its turned out to be a really nice CPU, but with that said, if I was building my PC now. I would spend the extra 50 bucks and bought the i7 10700K.
I really hope with AMD next CPU line. That there is a real performance different between there CPU line-up. I feel like you can spend 160 dollars on the 3600 and get a ver comparable gaming experience vs 3700x/3800x (xt)/3900x(xt). I feel that the advantage you get out of the other Ryzen CPUs are the games that like more cores/threads.
reply
I ended up buying a Ryzen 3700x over Intel at the time.
1) I was upgrading from i5 6600k (stock) and noticed with games like RDR2 - it did not respect 4 cores 4/4 and was having a unsmooth experience with some games.
2) Intel 9th Gen series was only available at the time and they dropped hypertherding from there i5/i7 series. And reviews seems to be suggesting that they where not as smooth becuase the lack of hypertherding.
So I took a chance on a AMD build and it has been okay, no real complaints. Its turned out to be a really nice CPU, but with that said, if I was building my PC now. I would spend the extra 50 bucks and bought the i7 10700K.
I really hope with AMD next CPU line. That there is a real performance different between there CPU line-up. I feel like you can spend 160 dollars on the 3600 and get a ver comparable gaming experience vs 3700x/3800x (xt)/3900x(xt). I feel that the advantage you get out of the other Ryzen CPUs are the games that like more cores/threads.
reply
Speejays
I'm not an AMD fanboy (Navi's drivers are still shitty a year later, Ryzen 3000 CPUs had some really shitty BIOS problems at launch, and Intel 10th gen is honestly a great value compared to Ryzen atm), but I think it's kind of unfair to show those Ryzen is smoother comments which were made before Intel's 10th gen came out and then do frametime comparisons using their 10th gen CPUs.
Isn't the point of those comments about how Ryzen has had 6 cores/12 threads for 200 and 8 cores/16 threads for 300 for 3 generations while Intel has had nowhere near that for their 7th gen/8th gen/9th gen CPUs? Intel responded back with their 10th gen matching Ryzen 3000, and now it's not really an argument anymore, but it seemed to have been a valid point for the past 3 years.
Imo, you should've tested using the past 3 CPU generations from both companies (including Ryzen 5s and Core i5s), because I'm still not really certain if those comments were correct or not during that time.
reply
I'm not an AMD fanboy (Navi's drivers are still shitty a year later, Ryzen 3000 CPUs had some really shitty BIOS problems at launch, and Intel 10th gen is honestly a great value compared to Ryzen atm), but I think it's kind of unfair to show those Ryzen is smoother comments which were made before Intel's 10th gen came out and then do frametime comparisons using their 10th gen CPUs.
Isn't the point of those comments about how Ryzen has had 6 cores/12 threads for 200 and 8 cores/16 threads for 300 for 3 generations while Intel has had nowhere near that for their 7th gen/8th gen/9th gen CPUs? Intel responded back with their 10th gen matching Ryzen 3000, and now it's not really an argument anymore, but it seemed to have been a valid point for the past 3 years.
Imo, you should've tested using the past 3 CPU generations from both companies (including Ryzen 5s and Core i5s), because I'm still not really certain if those comments were correct or not during that time.
reply
Idaefix
I am using a 6600K and a RX480 and I'm experiencing stutters on most games. Can someone hook me up with a method for troubleshooting it ? I can only make wild guesses on the numbers I'm producing.
I have been desperate to figure out which of those two is pulling it down. I've reviewed both usage and ofc the graphics card is basically always maxed out, what is more interesting is that the cpu hovers from 30 to 60 depending on the game, with regular spikes to 80 86% CPU usage. Is it possible that going for a more modern CPU (I don't care what brand I've been using both to great succes for years) would improve my experience ? Or should that 6600K (with a simple all core 4.2ghz overclock tho) still be able to handle those things ?
I've noticed more and more games are actually recommending Ryzen/Core 7xxx + as MINIMAL requirements, while the 480 is mostly still on the green side.
reply
I am using a 6600K and a RX480 and I'm experiencing stutters on most games. Can someone hook me up with a method for troubleshooting it ? I can only make wild guesses on the numbers I'm producing.
I have been desperate to figure out which of those two is pulling it down. I've reviewed both usage and ofc the graphics card is basically always maxed out, what is more interesting is that the cpu hovers from 30 to 60 depending on the game, with regular spikes to 80 86% CPU usage. Is it possible that going for a more modern CPU (I don't care what brand I've been using both to great succes for years) would improve my experience ? Or should that 6600K (with a simple all core 4.2ghz overclock tho) still be able to handle those things ?
I've noticed more and more games are actually recommending Ryzen/Core 7xxx + as MINIMAL requirements, while the 480 is mostly still on the green side.
reply
BruceLeedar
Setting aside the pissing contest, I do think these days that CPU benchmarks with cutting edge games should do these frame time graphs, rather than simply avg and 1%/0.1% lows, since games are increasingly depending on multiple core, potentially leading to inconsistent frame intervals if compensating for inadequate available cores to use.
Also, what of other 'smoothness' metrics unrelated to frame generation, e.g. input responsiveness? I don't know if this is a thing, but it seems possible that modern games might de-couple their visual rendering threads from the game world thread, potentially leading a situation where frame times are consistent (if the rendering threads have priority in a core-starved environment), but yet there is a lag in the functional game world.
reply
Setting aside the pissing contest, I do think these days that CPU benchmarks with cutting edge games should do these frame time graphs, rather than simply avg and 1%/0.1% lows, since games are increasingly depending on multiple core, potentially leading to inconsistent frame intervals if compensating for inadequate available cores to use.
Also, what of other 'smoothness' metrics unrelated to frame generation, e.g. input responsiveness? I don't know if this is a thing, but it seems possible that modern games might de-couple their visual rendering threads from the game world thread, potentially leading a situation where frame times are consistent (if the rendering threads have priority in a core-starved environment), but yet there is a lag in the functional game world.
reply
Francois
just out of curiosity... why do reviewers use 3200mhz cl14 memory when benchmarking AMD vs intel and then say its equal ?
why not use decent 3600mhz cl14 on both intel and AMD and call it equal ? because it seems like they are deliberately handicapping AMD, amd is memory sensitive, intel is not so which begs the question as to why ?
And to add before i get chewed out, the vast majority of ram is optimized for intel with XMP and very few are for AMD with AMP, so shoving intel optimized ram into a AMD build and calling it equal is well to put it blunt, not equal...
reply
just out of curiosity... why do reviewers use 3200mhz cl14 memory when benchmarking AMD vs intel and then say its equal ?
why not use decent 3600mhz cl14 on both intel and AMD and call it equal ? because it seems like they are deliberately handicapping AMD, amd is memory sensitive, intel is not so which begs the question as to why ?
And to add before i get chewed out, the vast majority of ram is optimized for intel with XMP and very few are for AMD with AMP, so shoving intel optimized ram into a AMD build and calling it equal is well to put it blunt, not equal...
reply
Add a review, comment
Other channel videos
