Hi, this is Wayne again with a topic “Overclocking — Awesome or Useless? — 3.5GHz vs 4.5GHz in CPU Bound Gameplay”.
Hello and welcome to a special bonus edition of stock versus overclocked. Thank you to those of you who support on floatplane in youtube. This video is being made available, at least when it’s published as an exclusive to supporters of the channel. Now I’ve done several videos on this configuration. I’Ve done several videos on this build there’s a playlist down in the video description below, and I’ve done, Ghost Recon break point comparing stock versus overclocked. But this time we’re taking a look at shadow of the Tomb Raider in the built-in benchmark.
Now in a built-in benchmark, you get an exact one-to-one copy of. What’S going on on the screen, you don’t have the variable gameplay that you get like you got in the Ghost Recon break point because I didn’t use the built-in benchmark. I showed live gameplay. I think it’s more illustrative of what you can expect in the real world, but it’s not exactly directly comparable, so we do have built-in benchmarks to provide that as well. A combination of the two, I think is ideal.
On the left hand, side we have stock clock speed at 3.5 gigahertz and on the right hand, side. We have overclocked at 4.5 gigahertz, so the 34 multiplier 133 megahertz base clock, which is the stock speed of an i9. Excuse me an i7 980 X now the X 56 75 was about 190 base. Clock overclock would give you very very similar performance to that and it’s frankly what most people should do as opposed to spending what an i7 980 X cost but well.
This was just much easier for me to test one point of note: the RAM speed numbers on the screen are in error. I, what can I say my mind – was just somewhere else when I typed that into MSI Afterburner we have 12 gigabytes of ddr3, 1600 and that’ll be reflected in the chart that you see later, it’s not 1333 at 1600, all the numbers on Emma’s afterburner. You see on the screen, those are simply typed into afterburner. You could type anything you want.
So it’s not like it’s reading like your Rams, speed or or how much beer am you have or what not it’s. Just you type, those into MSI Afterburner. I do have to apologize for the fact that I’ve got a graph on the right-hand side and not on the left. These were tested on different days at different times after I’ve messed around with the overlay. Yes, it would be nice to go back and retest everything and make it match exactly. What can I say, I’m human, in fact, the graph on the right is actually the frame rate graph, not the frame time graph going forward.
My plan is to standardize with taking the millisecond count off of the number. So, like the white number, you see there show in the millisecond count and take that off and the graph below is going to be the frame time graph and so it’ll just be real time. Average running average one percent low running one. Well, I shouldn’t say average, but the running one percent low and then the running point one percent loan.
This provides four digits. The other thing you’ll notice, is that on the left side in the stock configuration, we have five digits on the GPU. This is also something I’m going to standardize going forward. The percentage number in the middle is actually the percentage fan speed. The percentage number on the left is the graphics card usage and I think it’s confusing to have two percentages there. Some taking I’m standardizing it to only have four digits one which will let me make the text bigger and make it easier for you guys to see on smaller displays on the left-hand side.
You notice the text is a bit smaller and in order to get five columns of numbers in the top line, for example, you’ve got GPU usage, GPU temperature fan percentage, fan, rpm and then, of course, core clock of the GPU. But on the right-hand side, I’ve taken off the fan percentage, but it didn’t take off the frame times in terms of a number and then the graph is the frame rate rather than frame times. These benchmarks were actually not done like last week. I’M voicing this over March 22nd 2020. These benchmarks were actually recorded December 11. Yes, these were done quite some time ago.
All of these videos and all the content on these machines was supposed to have been done back and December. This is the whole video editor mess that I had in November. In December, we went through five different video editors trying to get this stuff done. In fact, the two main videos and one bonus, video which turned out to be the Ghost Recon breakpoint video, was supposed to have been done by a vo at her actually sent all of this material, the raw files, the charts and everything to two different editors and Never got anything back from either one of them that was useful.
One was way late. One sent me a demo project that wasn’t very good and then he didn’t follow up for two weeks, which is why this kind of got shelved, because that other stuff I had to do in January and I’m finally, in March, coming back around to this but yeah. This this was done some time ago. So that’s one of the reasons why you know.
I’Ve recently talked about the fact that I’m no longer doing 0.1 % lows. I’M gon na leave them in there just because I think there will be occasionally times people care about them and by taking off the frame time graphs on the white line numbers, then you can have four numbers and it fits just fine, because I thought about leaving The millisecond count up there and then taking off the point, one percent low. But if you take off the one point the point one percent low, then I don’t have it to give you numbers so well, the joys of benchmarking. I haven’t really talked about the actual numbers on the screen very much they’re self-evident, and you can certainly see them.
It is faster overclocked. It is not dramatically faster, overclock, it’s a little bit, but it’s not as much as you might think. 85 frames per second average at stock speeds, 94 frames per second, a 1 gigahertz overclock. That is a maybe 10 percent performance improvement for, like a 25 plus percent clock, speed improvement.
Remember that when you overclock your qpi link, your DDR Ram, your PCI Express Gen 2 bus. None of it goes faster, nothing improves besides the raw clock, speed of the CPU and because nothing else is getting faster. It just doesn’t make the difference. You might think it does unless you’re doing a truly cpu-bound test and you’re, not using your RAM you’re, not using your qpi links to your storage you’re, not using your PCI Express Gen to bus if you’re, not using any.
Of that then great I mean you’ll get a performance boost bottlenecks prevents you from running faster. Only when their bottlenecks I mean. The bottleneck here is not necessarily the CPU while it is, but it’s also the RAM and the bus speed and the PCI Express gen3 Gen. 2 bus, those things are all impacting performance in ways that just taking the clock, speed up alone doesn’t actually make that big of a difference on, and I’ve now done this on several different games, and it really doesn’t make a whole lot of difference, even though the Graphics card was kind of bored now, moving right along.
I’Ve got a bonus bit for you. What happens if we compare a Rison, 720 700 X with an RT X 2080? Not a super, and not a TI at stock. Clock speeds with some very boring. Ddr4.
3000 CL 16, not even the superfast to anything, is just very basic generic Ram. That’S not special, and we compare that to the i7 980 X with an RT X, 20 atti overclocked again. The numbers on the screen are wrong. Its ddr3, 1600 C online, I believe, is what that is, and it’s triple channel, so it’s faster than you would think, because this triple channel where’s, the ddr4, is double channel.
I mean that doesn’t make up all the difference, but it helps now for those of you playing the home game. You could look at the numbers on the screen and go oh well see here. We’Ve got a current running average of a hundred and three on the left and 95 on the right.
But the i7 980 X is powered by an RT X, 20 atti, which is over $ 1000, at least as I’m voicing this over versus an RT X 2080, which you can currently buy for $ 600. In fact, if you have an i7 980 980 xi7 920, you can buy a rise in font of 2700 xcp. You a decent B 450 motherboard 16 gigs of DDR 432 100 megahertz ram and an RT x 2080 for less total you’re gon na spend about $ 900 on that CPU motherboard, RAM and graphics card you’ll spend about $ 200 less on the configuration on the left. That’S everything you have to replace coz the RAM changed. Obviously you can keep your storage. You can keep your case. You can keep a power supply, then just putting a 20 80 TI on a ten-year-old. Now, please, don’t put 10. You know thousand our graphics cards on 10 year old machines.
That’S absurd! If you have 20 80 TI money, you have Zen to money. You have I 999 hundred K money. You don’t need to be doing something silly, like 20 ATT eyes, but the point is to lift the graphics card is a limit to test pure CPU performance, we’re doing theoretical testing here to some extent, how much does the CPU really affect the performance and the fact Of the matter is a Zen second-generation Rison, 720 700 X with a lower level. Graphics card is faster because it’s the total package and the 28 ATI is being held back by the i7 980 X and frankly, general Windows, compute performance general, launching the game, opening stuff running Windows, Update just using Chrome watching youtube videos watching twitch streams. The rise in 720 700 X is legit a much better experience than a 10 year old platform because of the RAM and the qpi link and the PCI Express Gen 2 bus and love black event via me, and a variety of other things. 94 frames per second on the 4.5 gigahertz overclock 103 on a stock 2700 X with a lower class GPU, with some very cheap basic Ram.
Nothing fancy about it same thing with the 1 % low 56 to 265 is a legit improvement. Now, on the point one percent low – it’s a built-in bench mark – don’t stress too much – I mean you could run out six times to get a different number, because it’s the it’s not a long enough test for the 0.1 % low to mean anything. Thank you all.
So much for watching this special bonus, video that is being published early to our YouTube members and to our floatplane subscribers. We greatly appreciate the support that you offer us normally I’d say like this video and subscribe and all that below. But you guys are our supporters.
So you guys already do that and we greatly greatly appreciate it thanks so much for watching this. Let me know what you think down below and I will see all of you next time. You .