• Welcome to TechPowerUp Forums, Guest! Please check out our forum guidelines for info related to our community.
  • The forums have been upgraded with support for dark mode. By default it will follow the setting on your system/browser. You may override it by scrolling to the end of the page and clicking the gears icon.

How high of a ram frequency can i run on a Z690 with an 14700Kf processor?

I don't think it's entirely not doable, but this is a 24/7 setting at an easy to manage voltage, I don't need active cooling or anything. Really, not much of a bother.
I daily 8600MT's, doesn't need an active fan for normal use case..stable for months now since I was able to do it..
 
Until you do upgrade bump your virtual memory page file to at least meet the x1.5 recommendation or even more more x2.0 or x3.0 it might help in your situation when it's trying to juggle a remaining like 8GB of system memory beyond what the GPU is trying to hog to keep itself fed with that 24GB of VRAM. You'd suffice with 48GB I think well enough it's still double what the GPU has which pretty much fine and those kits are cheaper and scale better anyway. I definitely agree more than 32GB makes sense in your case especially, but as far as do 64GB or would 48GB suffice quite and in general not really hold you back probably, but there are always outlier cases and obviously if you multi-task a lot and run plenty in the background more than 64GB will quickly more more sense. I still think in about 98% of the situations for the duration you'd use that particular hardware setup of CPU/MB/GPU you wouldn't notice outside like if you're doing AI stuff and needs job memory for training which is really for training as opposed to running a algorithm that's trained.
 
Last edited:
I daily 8600MT's, doesn't need an active fan for normal use case..stable for months now since I was able to do it..
I don't think it's that simple to say 8600 MT/s is good without a fan for normal use case. What is "normal". Also if you change the sub-timings, a fan is a must I think. Errors be upon you at 45c.
 
I don't think it's that simple to say 8600 MT/s is good without a fan for normal use case. What is "normal". Also if you change the sub-timings, a fan is a must I think. Errors be upon you at 45c.
I never reach 40c even with hours of gaming..ever heard of the Lian Li V3000+ case and 15 120mm fans??

I have seen y-cruncher rape my sticks beyond 45c and never had stability issues, there those who say DDR5 is too sensitive to temps, though I have seen rather "more extreme" on other forums and if its really stable it won't give a puke about errors even beyond 45c..

Normal for me is gaming and reading/watching some stuff over the internet, that's what I do with the big boi of my rigs..my bench rig is either a 2nd Apex Encore with another i9/i5 or the Gigabyte Aorus Tachyon B650E with a turd of a chip 9700X..
 
Last edited:
I never reach 40c even with hours of gaming..ever heard of the Lian Li V3000+ case and 15 120mm fans??
Exactly gaming.. You can run pretty unstable settings and be okay for gaming. Minus the random shuddering that comes from bad calculations. I also have that case, so what?

I have seen y-cruncher rape my sticks beyond 45c and never had stability issues, there those who say DDR5 is too sensitive to temps, though I have seen rather "more extreme" on other forums and if its really stable it won't give a puke about errors even beyond 45c..
I've done enough testing and data collection to definitely say DDR5 is quite sensitive to temperature....at least for SK Hyinx that is. Not enough samples to figure out Micron or Samsung yet. But If your using a kit with SK Hynix, straight up XMP will error out in the 60s range. Mess with the tREFI and other sub-timings and you are down in the low 40s. Of course both CAS and Frequency changes, so I am just giving a ballpark number.

The next Thermal Testing article I do will have a massive section of CAS vs Temp. Just takes 100+ hours to gather the data for a single chart... Probably another 6 months before I finish part 2.
 
Exactly gaming.. You can run pretty unstable settings and be okay for gaming. Minus the random shuddering that comes from bad calculations. I also have that case, so what?
I passed Y-Cruncher VST, with 8600MT's, I need to bump my SA a bit..easy there, I mean no harm..silicon binning is a thing you know? so its not just skill/talent or what not, high octane voltages etc..etc..etc..
I've done enough testing and data collection to definitely say DDR5 is quite sensitive to temperature....at least for SK Hyinx that is. Not enough samples to figure out Micron or Samsung yet. But If your using a kit with SK Hynix, straight up XMP will error out in the 60s range. Mess with the tREFI and other sub-timings and you are down in the low 40s
i am running TREFI@131070, I use a special fan for the memory, although those are only used on the bench jig, on the "Normal Use" PC, I have an active cooling solution blowing cold air on the face of the Memory sticks, I also thermal puttied my sticks..spending 300$ on the iceman waterblock for RAM doesn't sound practical personally at all, I am no sponsored sheep or what not..I fund the hobby by my own pocket..
 
Tell tale sign of a mad cow when it starts making sponsored sheep sounds...:laugh: There is more often than not going to be a fair bit of stability variance in all directions when it comes to running bleeding edge technology on newer emerging hardware running it. Someone with Apex with a top tier CPU might have a piece of cake time at it, but someone one a mid tier CPU and previous generation MB of lower quality even with the same bin quality memory is probably not going to have as easy a time since more than one variable factors into the equation. It runs for me on top quality hardware piece of cake this is everyone's experience no it really isn't sorry be realistic. I think it's certainly getting more approachable to run higher binned memory even on some older hardware in cases, but don't expect it to be smooth sailing for everyone. You can sometimes deduct reasonable assumptions and roll the dice a bit, but it's not really something to just generally recommend people do so.
 
I find its really not worth it. I've not had high end ddr5, my ram is also 5600, cl36. But it is very overclockable. And I can get 6800 ( with high latencies) or 6400 cl30 out of it. But I choose to run 6000 cl32 ( with a few other tweaked timings) because I can get very low voltages on the cpu side (lower than normal xmp, I can run SA at 1.15 instead of xmp default that was 1.25 I think, and vddq_tx at 1.2 instead of 1.3 default xmp or 1.4 - which is what I need for those higher clocks) and just slightly higher voltage on the ram side, (1.25 > 1.35 - need 1.5 for 6400 cl30 or 6800). In terms of performance it really doesn't make a difference when you game at 4k. Stability is more important than numbers being high for the sake of being high imo.

Perhaps there's some scenarios where it would be worthwhile, some productivity work that relies heavily on ram, or perhaps some games at low res.... But me, I'm gonna stick with 6000. I did do the 6400 cl30 for quite a while, idk a year perhaps, so the ram is more than capable of handling the voltage, at least medium term, but still, I don't see the point really anymore, outside of benchmarks, which is the only time I'll ramp them up.

Having my pc work for a long time is of a lot of importance to me which is why I don't push it. At least not anymore.... took a failure for me to think differently about things.

As for theoretically whats the highest frequency you could get, it really depends on combination of cpu + ram die ( hynix is the best, and you want single rank) + motherboard (high end two slot is the best) and of course the silicon lottery. So, hard to say. There's a lot of luck to it. You'd have to experiment to find out for yourself.
 
Last edited:
You are implying I am a sponsored sheep? baa baa baa baa baa baa baa baa baa baa baa baa baa baa baa baa baa
I did never mentioned/dropped a name? don't worry just now, I ran my 8200CL36 profile so you can sit on top of whatever you're implying as a throne or ranks..I'm gonna be the good guy here and run slow.. :peace:
 
I did never mentioned/dropped a name? don't worry just now, I ran my 8200CL36 profile so you can sit on top of whatever you're implying as a throne or ranks..I'm gonna be the good guy here and run slow.. :peace:
I'm fine at 6800 CL28. Mid tier board and memory kit.

If ir_cow reviews hardware, then he gets a free trial run. Little be known to a lot of people, some venders request the hardware returned. Just wanted you to know that...
 
I'm fine at 6800 CL28. Mid tier board and memory kit.

If ir_cow reviews hardware, then he gets a free trial run. Little be known to a lot of people, some venders request the hardware returned. Just wanted you to know that...
I thought it was always free like a "gift" or token of appreciation for the upcoming unbiased review..anyways, I imagined things like HWUB with all that glorious hardware on the BG..
 
I thought it was always free like a "gift" or token of appreciation for the upcoming unbiased review..anyways, I imagined things like HWUB with all that glorious hardware on the BG..
Well, yeah, sometimes you get to keep the gear. It depends on a few things in the written agreements.

Maybe HWUB gets to keep most review samples. Also, some release samples aren't as good as the consumer product.

Is HWUB unbiased? I don't know. I wait till end product is released and look for consumer opinions most of the time. But if their review looks similar to everyone else's review, the information must be true.
 
Is HWUB unbiased? I don't know. I wait till end product is released and look for consumer opinions most of the time. But if their review looks similar to everyone else's review, the information must be true.
I never in particular said they were, pretty sure they're the opposite

I got a free Aorus B650E Tachyon and for all I know this board is still like a prototype of some sort of a degree, its a pain with choosing memory sticks, its picky and has tantrums on its own..probably due to its nature as a prototype..
 
I never in particular said they were, pretty sure they're the opposite

I got a free Aorus B650E Tachyon and for all I know this board is still like a prototype of some sort of a degree, its a pain with choosing memory sticks, its picky and has tantrums on its own..probably due to its nature as a prototype..
Or the fact it's a B series boards and a gigabyte. The 2 things it has against its self. :p
 
What would the settings be if i were to change the Expo I setting to Tweaked in my system? I'm running a 9800X3D, 32GB (2x16) 6000Mhz C28, with an Asus Prime X670E-Pro. Has anybody tried Tweaked settings instead of the XMP/EXPO settings before?
 
What would the settings be if i were to change the Expo I setting to Tweaked in my system? I'm running a 9800X3D, 32GB (2x16) 6000Mhz C28, with an Asus Prime X670E-Pro. Has anybody tried Tweaked settings instead of the XMP/EXPO settings before?
Rule of thumb concerning X3D chips specifically, the few changes in "tweaking" doesn't constitute the time and effort put in to obtain possible 1% realistic numbers.

You want stable forever, just run the XMP profile and enjoy. Want FPS, OC with the gpu instead. (At risk as always)
 
Back
Top