I have an i9 10900K and a 3080. Its the 1%, 0.1% lows that I am interested in when I fully tune my system. That's ram, core, ring and gpu. They all need to be done in order to get the full benefit. 20MB of cache just isn't cutting it these days.
love this video and I really respect the amount of work you put into it. really did not expect a full overclock to perform that well. also, the conclusion is based around the idea that 1% lows have a stronger impact on how the game feels than average FPS. I totally agree, and hope that more people take note of this.
It's criminal that this video only has 628 views. This is an incredible amount of useful, practical information, on hardware that is "common." You explain your methodology perfectly and summarize the data very well. Fantastic video, and I'm glad to see there's a reason to spend as much time as I do overclocking!
Good job with the data here. Biggest complaint id say is about a static OC. Its actually faster. Single core boost matters very little. Static oc gives much better 1% and .1% lows. Test it out for yourself
Yeah i was pretty annoyed with jayz2cents having yet another vid about oc being dead... while he clearly lost every grip on reality with his i9s. He just uses the wrong hardware to oc. You can be happy if you can cool the i9 at stock speeds which is pretty impossible on air or water. And ryzen basically is sold oced. Take the i5 12600k and it will oc like a champ.
@@CmdrSoyo Right. Got the i5 at 5.4 ghz turbo and 5.3 all core, 4.3 ghz on the baby cores with a 240mm aio, which is hardly better then a high end air cooler, never go above 80 degrees core temp. Take a oc on every other part of the pc and yes it makes a world of difference.
@@CmdrSoyo 13700K is a fun CPU. It's amazing how the mainstream just abandoned having any fun with the hardware anymore. Everyone just looks at the the extreme top end on tech tube. It really is cementing the false narratives we see now. Had a dude buy a 13900K and a 4070ti for 4k gaming. He had been inundated that simply buying the top would be all he needs, and that DLSS3 would be all you need..... It's truly amazing how angry and defensive one gets when you tell them they could have saved hundreds of dollars for the same performance.
Just tested HZD on my old tuned R5 5600X setup vs my new R5 7600 at stock with ram at jdec 4800.. the new cpu was 25% better in the 1% lows, 27% in average.. running the original quality preset at 1080p. so the new cpu at stock was almost "2 generations" faster then the overclocked previous gen in my case
If you did this test with a 5600X instead of a 5800X the pbo gains would be considerably bigger, because the stock clock and power limit on that is only 4650mhz / 73W.
Going through the video for the longest of time and couldn't find details of what resolution and graphics details are you using. So - what resolution and graphics setting are you using?
i still wonder why you care though? unless you have the exact same hardware and want to use the numbers as a guide/goal line knowing the settings is useless to you.
@@CmdrSoyo it's to understand full settings of the condition so I can appreciate your numbers better. And yes correct on that part, it is useless unless we have same hardware. Nevertheless I appreciate your time and effort, it's good information. Thank you.
flat primaries at high frequency are an indicator of B-Die but not a requirement. similar to how i got H8D in an identical spec kit and ran DDR4-5000 on it.
I wish I was smart enough to understand mem tuning. I currently run my 7800X3D ram at lower than expected speeds after attempting to manually tune them. Negative PBO though is easy to stress test and has a clear benefit on noise levels (without PBO this CPU set itself on fire when anything is using it at all, after -25 PBO it's still cool at 60% util) and boosting to max freq all the time now. Very interesting tests, reducing stutter is important for games and for windows reactivity, so that really shows how important it is to do the work. But mem tuning is just too hard I don't understand half of the settings haha. If anyone has a real guide about memory tuning it be great..
yeah i have a friend with a 12900K and 3600 16 19 19 xmp and I run better fps than him on overwatch with a 8700K 4133 15 15 15 with manual sub and terts
Mainstream points to OC being dead on a chip where it's no longer appreciably scaling with just pumping clocks. I think only HUB really has a 7200mts with some tightened timings rig for their intel testing. They pushed how good scaling for DDR4 with B die could be under the rug, and they didn't add Faster DDR5 until the disparity between speeds on Zen 4 and Intel were so massive to ignore... and even then much of the benefit is from latency reduction which is part and parcel of timings tuning..... Which is fine, no problem.... but when the consensus become "OC is useless", it's disingenuous. I mean, Jay's last vid with memory was not properly done, and his kits were not stable. It contributes to people making a conclusion on bad info. I tell people not to bother with OC if they don't want to devote the time and effort. I say the performance gains exist, but it's the will to make mistakes and gain knowledge of their hardware that is required. If that's not an enjoyable thing, don't worry and just play the games, but don't complain about $$ and needing to upgrade when you haven't touched the hardware yet.
I have an i9 10900K and a 3080. Its the 1%, 0.1% lows that I am interested in when I fully tune my system. That's ram, core, ring and gpu. They all need to be done in order to get the full benefit. 20MB of cache just isn't cutting it these days.
A monumental video! Absolutely wonderful video putting to words what I’ve been saying for the last couple years.
love this video and I really respect the amount of work you put into it. really did not expect a full overclock to perform that well.
also, the conclusion is based around the idea that 1% lows have a stronger impact on how the game feels than average FPS. I totally agree, and hope that more people take note of this.
Thank you Soyo !
Awesome stuff fella!
It's criminal that this video only has 628 views. This is an incredible amount of useful, practical information, on hardware that is "common." You explain your methodology perfectly and summarize the data very well. Fantastic video, and I'm glad to see there's a reason to spend as much time as I do overclocking!
there is nothing "common" about B-Die RAM and 1080ti in 2024
Wait there only sub 1000
Still
Good job with the data here. Biggest complaint id say is about a static OC. Its actually faster. Single core boost matters very little. Static oc gives much better 1% and .1% lows. Test it out for yourself
Great video! How long did it take you to fully tune the ram? is it 100% stable?
It's based off previous overclocks i had done so it was pretty quick. Maybe a combined 2 days. It passed 1000% memtest stable
Yeah i was pretty annoyed with jayz2cents having yet another vid about oc being dead... while he clearly lost every grip on reality with his i9s. He just uses the wrong hardware to oc. You can be happy if you can cool the i9 at stock speeds which is pretty impossible on air or water. And ryzen basically is sold oced. Take the i5 12600k and it will oc like a champ.
I built a system with a 13700K a while ago for a friend and he had a single fan air cooler and it still got to 5.6GHz all core lol
@@CmdrSoyo Right. Got the i5 at 5.4 ghz turbo and 5.3 all core, 4.3 ghz on the baby cores with a 240mm aio, which is hardly better then a high end air cooler, never go above 80 degrees core temp. Take a oc on every other part of the pc and yes it makes a world of difference.
@@CmdrSoyo 13700K is a fun CPU. It's amazing how the mainstream just abandoned having any fun with the hardware anymore. Everyone just looks at the the extreme top end on tech tube. It really is cementing the false narratives we see now.
Had a dude buy a 13900K and a 4070ti for 4k gaming. He had been inundated that simply buying the top would be all he needs, and that DLSS3 would be all you need..... It's truly amazing how angry and defensive one gets when you tell them they could have saved hundreds of dollars for the same performance.
Just tested HZD on my old tuned R5 5600X setup vs my new R5 7600 at stock with ram at jdec 4800.. the new cpu was 25% better in the 1% lows, 27% in average.. running the original quality preset at 1080p.
so the new cpu at stock was almost "2 generations" faster then the overclocked previous gen in my case
If you did this test with a 5600X instead of a 5800X the pbo gains would be considerably bigger, because the stock clock and power limit on that is only 4650mhz / 73W.
Nice video!!
Is that kit actually bdie? Never seen bdie that slow
i have a i9 10850k but i am planning on making an amd mini itx build and sell my current build.
Going through the video for the longest of time and couldn't find details of what resolution and graphics details are you using.
So - what resolution and graphics setting are you using?
Whatever gave the best quality at the highest FPS like i said in the video.
@@CmdrSoyo finally heard it at 12min+ mark. Only mentioned medium to high settings, no mention of resolution.
Anyway, thanks.
@@goldfries it's all 1080p. i disabled any upscaling when available.
i still wonder why you care though? unless you have the exact same hardware and want to use the numbers as a guide/goal line knowing the settings is useless to you.
@@CmdrSoyo it's to understand full settings of the condition so I can appreciate your numbers better.
And yes correct on that part, it is useless unless we have same hardware. Nevertheless I appreciate your time and effort, it's good information.
Thank you.
Great video, can you show how to manually oc the ram? for example i have g skill 3600 16 19 19 39 i want more tight timing.
Hey, is that ram really b die? According to what I’ve been told on Reddit it always comes with the same number in sequence like (16-16-16-16)
flat primaries at high frequency are an indicator of B-Die but not a requirement. similar to how i got H8D in an identical spec kit and ran DDR4-5000 on it.
I wish I was smart enough to understand mem tuning. I currently run my 7800X3D ram at lower than expected speeds after attempting to manually tune them. Negative PBO though is easy to stress test and has a clear benefit on noise levels (without PBO this CPU set itself on fire when anything is using it at all, after -25 PBO it's still cool at 60% util) and boosting to max freq all the time now.
Very interesting tests, reducing stutter is important for games and for windows reactivity, so that really shows how important it is to do the work. But mem tuning is just too hard I don't understand half of the settings haha. If anyone has a real guide about memory tuning it be great..
yeah i have a friend with a 12900K and 3600 16 19 19 xmp and I run better fps than him on overwatch with a 8700K 4133 15 15 15 with manual sub and terts
Mainstream points to OC being dead on a chip where it's no longer appreciably scaling with just pumping clocks. I think only HUB really has a 7200mts with some tightened timings rig for their intel testing. They pushed how good scaling for DDR4 with B die could be under the rug, and they didn't add Faster DDR5 until the disparity between speeds on Zen 4 and Intel were so massive to ignore... and even then much of the benefit is from latency reduction which is part and parcel of timings tuning..... Which is fine, no problem.... but when the consensus become "OC is useless", it's disingenuous. I mean, Jay's last vid with memory was not properly done, and his kits were not stable.
It contributes to people making a conclusion on bad info. I tell people not to bother with OC if they don't want to devote the time and effort. I say the performance gains exist, but it's the will to make mistakes and gain knowledge of their hardware that is required. If that's not an enjoyable thing, don't worry and just play the games, but don't complain about $$ and needing to upgrade when you haven't touched the hardware yet.