Gadgetory

hey guys welcome back to another harbor box video I'm not quite sure how to properly introduce this video because in a way it's a bit pointless or at least the testings a bit pointless and I know that's not a great way to start off but I found the results interesting and I thought you guys might as well basically over the past few weeks in my spare time something I sadly have very little at the moment I've been madly benchmarking CPUs the primary reason for this is to build up a database of results for when Caddy Lake and then finally arrive who's then every time I talk about Zen I get excited anyway when benchmarking intel's old Sandy Ridge processes I noticed something strange they were much slower than expected not that long ago I did a huge amount of testing for a video map presented which looked at how the last five generations of intel core i7 processors compared in terms of performance when gaming with all five processors clocked at four gigahertz the Sandy Bridge 2600 K was for the most part not a great deal slower than the 6700 K at the most extreme when looking at the minimum frame rate we found the 2600 K to be around 20% slower this is a reasonably large drop-off in performance though given the age of the Sandy Bridge architecture we didn't feel this was significant moving forward to the here and now when making a similar comparison between these processes I found the 2600 K was at times more than 30 percent slower that's a pretty big difference compared to what was found around nine months ago so I wanted to find out what was going on was I doing something wrong or was there some other factor at play the key difference between now and then other than the games being tested which no doubt contributes to some of the performance hit I'm now using a significantly more powerful GPU previously testing was done with the 980ti but now I'm testing with the more updated Titan XP for those wondering the reason why we use such a fast GPU or at least the fastest GPU available time of testing is to remove this component as a performance limiting factor and this places more emphasis on CPU performance so was the combination of new or more demanding games plus a more powerful GPU their only reason why Sandy Bridge fell further behind as it turns out there's another factor to consider the PCI Express bus as many of you know Sandy Bridge processors support PCI Express 2.0 rather than the current 3.0 revision this means the link between the PCIe x 16 slot and the cpu is limited to 8 gigabytes per second on the other hand Ivy Bridge in newer processes offer 16 gigabytes per second as they use the 3.0 revision in the past this hasn't really been a big deal and we found that with previous generation gtx 980ti and r9 fury x flagship GPUs that it only cost gamers around a frame or two at most although it hadn't been an issue in the past was the limited PCI Express bound with an issue for the Titan XP thankfully there was an easy way to find out with my asrock Zed 170 motherboard which allows me to change the PCIe x 16 mode from 3.0 to 2.0 and even 1.0 I decided to test all three revisions with the Titan XP as well as the GTX 1080 and the gtx 1070 graphics cards and yes I know I know we are such Nvidia shells we tested with no AMD cards once again but as I said previously the fury X we've already know it doesn't benefit from the latest revision so there isn't much point testing again and the sames obviously going to be true for the RX 480 which is slower again and on that note the GTX 1060 it's not going to benefit because we know that's around gtx 980 performance and previously we didn't see much difference with the 980 TI anyway circling back the reason why i said the testing was pointless was because no one is ever going to find themselves running a gtx 1080 let alone a titan XP on a platform that doesn't support the latest pci express revision i suppose if they did though it would look something like that titan XP on a zd68 motherboard with a 2600 k processor yeah I don't think so either still I've gathered all the results for my own investigation to find out what was going on and I found them interesting thought you guys might as well and decided to share them so let's check them out testing with Gears of War 4 you can see that when using the gtx 1070 there is no difference in performance between the second and third generation PCIe revisions in fact there's almost difference when looking at the very first revision either the gtx 1080 also provides similar performance trends for the most part moving up to the titan XP things start to change here we see a 5 FPS jump in the minimum frame rate though this does only equate to a 5% increase but honestly that's larger than I was expecting the day use X mankind divider results are very interesting the gtx 1070 only sees a 1 to 2 FPS increase when upgrading from the second gen PCIe revision that said we do see a steady 3 FPS increase when coming from the first gen revision increasing the rendering power the gtx 1080 we find similar performance margins for the most part the same is true for the Titan XP results the margin widens by just a single frame the last game where around our full battery tests was battlefield one here the Titan XPS are a 4% increase for the minimum frame rate and a 6% increase for the average we're moving to the latest third generation revision that netted us eight FPS more on average not a game changer but it will make a difference for those running 144 Hertz monitors ok so the gains with the Titan XP weren't earth-shattering but we did see a three to five percent performance to climb when using the second generation PCI Express revision this goes some way to explaining the extra 10 percent Sandy Bridge losses to skylake in my recent tests the rest we put down to the Titan XP being much fast in the gtx 980ti and the fact that we're using more modern demanding no more modern games that are more demanding yeah more modern more demanding games I'm sure these results in below anyone's socks off but as I said I found them interesting and thought you guys might too there's not much else to say really you aren't going to run into any real-world scenario where a Titan XP is limited to pcie 2.0 or worse but now you know how to perform if it was on that bombshell I'm your host Steve catch you next time you