The RAM used in this video runs at 4800MT/s CL40 base speed, and at 6400MT/s CL32 with XMP enabled. Some games benefit more from higher speeds, some more from lower latency, and some from both.
This depends on the game. But lowering your latency will also lead to higher data transfer as there is less delay when the CPU needs to fetch something from RAM.
Thank you for this video. Just enabled XMP on system bringing ram from 2400mhz up to 3600mhz and was curious if that would make a difference in Jedi Survivor, so was looking for info on RAM speed in "open world games". Then you literally bench marked Jedi Survivor and showed that yes, it is a pretty dramatic difference lol. I did not want to install a 130 GB game for it just to not work. Thanks for the video! Much appreciated!
Since some games were running at 120+ frames so i wouldn't bother OCing anything at that point, maybe only undervolting. What is interesting is that in Jedi Survivor, when the avg went below 60 fps at some point OCing the RAM significantly improved performance. So maybe you could do the same testing with some cores disabled to emulate a budget CPU? I suspect that overclocking the RAM would have a bigger impact on a lesser system that is struggling to run the games. Anyway interesting content, I've just subscribed after watching your CPU bottleneck. - Cheers.
Really appreciate the sub! Yeah, on lesser CPUs the difference would be greater indeed. The more CPU limited you are the more faster RAM helps. And to show that difference people usually test with a 4090 at 720p for example, to enforce a major CPU bottleneck. I just wanted to see how much of a difference it actually makes in more real world scenarios, and to be honest I expected even less, as the 2nd set of settings used are generally very GPU bound. So I was quite surprised that there was a difference in the majority of the games here.
This is gpu test not a cpu. Most of settings in game heavily dependend on a cpu, so, to proper test cpu, you need use max|ultra setting in all games plus enable ray tracing(also love eat your cpu). In your test you lower settings(for some reason) and enable high resolution when gpu bound. In cyberpunk you need enable crowd to high.(super heavy on a cpu). Low settings in ghost of tsuhima, and after that you using 1440p. You almost always gpu bound. You can call it gpu test, not a cpu. Just enable 720p + ultra settings +ray tracing+dlss perf and see bigger difference between ram speed. with 720p + perfomace dlss you can simulate perfomance of 4090 for test purposes. fast ram + tuned timing = faster feeding cpu with data = more fps + lower fps dips. thanks for test.
Thanks for the comment. As I said in the video I tested more realistic settings. To show differences you'll use 720p, but I wanted to show what the differences would be using settings people actually use to play games with a similar kind of setup, and how much RAM speeds matter then.
@@Mostly_Positive_Reviews it just not enough perfomace of your 4070 to test cpu in high resolution. because of that this is not a clean ram test. this is a gpu test. And low settings not a "realistic setting" for sure. One day folks will upgrade their gpu to more powerfull and will not gain any fps boost because of cpu. You would have shown this. It is like in jedi max fps for cpu is ~80 but if you upgrade to 4090(double of 4070 perfomance) you are not gain more fps that 4070 had, because cpu can not do more draw calls. so logic behind this you will need upgrade and cpu to gain more fps.
Yes, and I have spoken about CPU binds in other videos where I say exactly this. This is not supposed to be a clean RAM test, as you call it, but a RAM test at realistic settings, that's it. These settings are what I use with this system, 1440p High with DLSS quality, and that's why I used those to test real world scenarios instead of just the usual 4090 at 720p to show theoretical differences.
It can be cpu dependant, some like high speeds some don't care. All depends on the cpu. Just look at the 5700x3d, runs a decent bit faster with 3600mhz cl16 rather than the usual 3200mhz cl16.
@@Mostly_Positive_Reviews Ram Overclocking is mostly depends on Motherboard only few z790 motherboards can run 7800mhz and above 2 dimm motherboard are the best for Ram Overclocking.
@@Decki777 Yeah, even 7200MT/s can be a struggle if the CPU's IMC or motherboard is just slightly not good enough. I think 6000MT/s is safe, and I have seen most people achieve least that. Hopefully that changes with the new CPUs and new platforms going forward.
@@Mostly_Positive_Reviews you are right. Good motherlands cost way too much and Ram tuning is hard and time consuming. ASRock z790i lighting is the best budget ram Overclocking it can run 8000mhz with tighter timings stable.
The RAM used in this video runs at 4800MT/s CL40 base speed, and at 6400MT/s CL32 with XMP enabled. Some games benefit more from higher speeds, some more from lower latency, and some from both.
What would you judge as more impactful, speed or latency?
I'm running cl40 right now, and have often wondered if its holding me back.
Just was well that was going to be my question as transfer rates is one thing latency is quite another.
This depends on the game. But lowering your latency will also lead to higher data transfer as there is less delay when the CPU needs to fetch something from RAM.
Would 40cl 5600mhz vs cl30 6000mhz make a big difference in counter strike 2?@@Mostly_Positive_Reviews
I ve always hated the fact we're supposed to OC the RAM in order to reach the max, promised speed.
Thank you for this video. Just enabled XMP on system bringing ram from 2400mhz up to 3600mhz and was curious if that would make a difference in Jedi Survivor, so was looking for info on RAM speed in "open world games". Then you literally bench marked Jedi Survivor and showed that yes, it is a pretty dramatic difference lol. I did not want to install a 130 GB game for it just to not work. Thanks for the video! Much appreciated!
Any time! Happy it helped you out!
Since some games were running at 120+ frames so i wouldn't bother OCing anything at that point, maybe only undervolting. What is interesting is that in Jedi Survivor, when the avg went below 60 fps at some point OCing the RAM significantly improved performance.
So maybe you could do the same testing with some cores disabled to emulate a budget CPU? I suspect that overclocking the RAM would have a bigger impact on a lesser system that is struggling to run the games.
Anyway interesting content, I've just subscribed after watching your CPU bottleneck. - Cheers.
Really appreciate the sub! Yeah, on lesser CPUs the difference would be greater indeed. The more CPU limited you are the more faster RAM helps. And to show that difference people usually test with a 4090 at 720p for example, to enforce a major CPU bottleneck.
I just wanted to see how much of a difference it actually makes in more real world scenarios, and to be honest I expected even less, as the 2nd set of settings used are generally very GPU bound. So I was quite surprised that there was a difference in the majority of the games here.
Wow, I wasnt drunk when I typed this, I promise. Just a phone with a screen protector...
This is gpu test not a cpu. Most of settings in game heavily dependend on a cpu, so, to proper test cpu, you need use max|ultra setting in all games plus enable ray tracing(also love eat your cpu). In your test you lower settings(for some reason) and enable high resolution when gpu bound. In cyberpunk you need enable crowd to high.(super heavy on a cpu). Low settings in ghost of tsuhima, and after that you using 1440p. You almost always gpu bound. You can call it gpu test, not a cpu. Just enable 720p + ultra settings +ray tracing+dlss perf and see bigger difference between ram speed. with 720p + perfomace dlss you can simulate perfomance of 4090 for test purposes. fast ram + tuned timing = faster feeding cpu with data = more fps + lower fps dips. thanks for test.
Thanks for the comment. As I said in the video I tested more realistic settings. To show differences you'll use 720p, but I wanted to show what the differences would be using settings people actually use to play games with a similar kind of setup, and how much RAM speeds matter then.
@@Mostly_Positive_Reviews it just not enough perfomace of your 4070 to test cpu in high resolution. because of that this is not a clean ram test. this is a gpu test. And low settings not a "realistic setting" for sure. One day folks will upgrade their gpu to more powerfull and will not gain any fps boost because of cpu. You would have shown this. It is like in jedi max fps for cpu is ~80 but if you upgrade to 4090(double of 4070 perfomance) you are not gain more fps that 4070 had, because cpu can not do more draw calls. so logic behind this you will need upgrade and cpu to gain more fps.
Yes, and I have spoken about CPU binds in other videos where I say exactly this. This is not supposed to be a clean RAM test, as you call it, but a RAM test at realistic settings, that's it. These settings are what I use with this system, 1440p High with DLSS quality, and that's why I used those to test real world scenarios instead of just the usual 4090 at 720p to show theoretical differences.
MPR my goat
Thank you! You're the real goat for sure!
It can be cpu dependant, some like high speeds some don't care. All depends on the cpu. Just look at the 5700x3d, runs a decent bit faster with 3600mhz cl16 rather than the usual 3200mhz cl16.
trying to catch up on what I missed!
Really appreciate it bud. Though I think you should save some of this money for backup power ;)
@@Mostly_Positive_Reviews hehe just a little something to help the channel
Thank you! You know I always appreciate this, a lot!
Ram speed and tight timings are matter not just speed
Yes, both matter. Some games like more speed, and some like lower latency.
@@Mostly_Positive_Reviews Ram Overclocking is mostly depends on Motherboard only few z790 motherboards can run 7800mhz and above 2 dimm motherboard are the best for Ram Overclocking.
@@Decki777 Yeah, even 7200MT/s can be a struggle if the CPU's IMC or motherboard is just slightly not good enough. I think 6000MT/s is safe, and I have seen most people achieve least that. Hopefully that changes with the new CPUs and new platforms going forward.
@@Mostly_Positive_Reviews you are right. Good motherlands cost way too much and Ram tuning is hard and time consuming. ASRock z790i lighting is the best budget ram Overclocking it can run 8000mhz with tighter timings stable.
Btw you make great videos @@Mostly_Positive_Reviews
its all about the lows