8 PACK MEMORY RANGE GROWING: SAY HELLO TO 8 PACK RIPPED EDITION & 32GB KITS!!!

[Nick1881]

I updated my bios today and now I can run FLCK 1900, I haven't tried any higher. I ran the same timings I previously ran when my 3700X worked at 1900, though I tuned back down then as it wouldn't always post.

How does this look? Anything else I should try tweaking?

 

[Guest2]

I updated my bios today and now I can run FLCK 1900, I haven't tried any higher. I ran the same timings I previously ran when my 3700X worked at 1900, though I tuned back down then as it wouldn't always post.

How does this look? Anything else I should try tweaking?

Try my CAS14 timings, 1.5V. Will lower latency a bit

https://www.overclockers.co.uk/forums/posts/34332163/

 

[Nick1881]

Try my CAS14 timings, 1.5V. Will lower latency a bit

https://www.overclockers.co.uk/forums/posts/34332163/

Tempting, I'm cautious of pushing the voltage higher though, I'm already at 1.47, though I haven't tried lower with this B550 board, just what it needed with B450.

Also this board can be a pain sometimes, I've had to clear cmos twice already after it refused to post with tiny changes, yet after clearing cmos and trying again it worked. Getting inside to clear cmos is an extra pain too with the underdesk mounting of my system.

I know that @MrPils looked at others and pointed out little tweaks that wouldn't hurt much. I keep reading his very informative posts but it's a bit much for my tiny brain at times

 

[BallistixOnZ490]

Tempting, I'm cautious of pushing the voltage higher though, I'm already at 1.47, though I haven't tried lower with this B550 board, just what it needed with B450.

Also this board can be a pain sometimes, I've had to clear cmos twice already after it refused to post with tiny changes, yet after clearing cmos and trying again it worked. Getting inside to clear cmos is an extra pain too with the underdesk mounting of my system.

I know that @MrPils looked at others and pointed out little tweaks that wouldn't hurt much. I keep reading his very informative posts but it's a bit much for my tiny brain at times

1.53v on my Micron E die only reaches 55c but for the surface heatspreader temps measured with an IR thermometer.

Both those and Samsung B die dont care much about voltage, 1.5v is nothing for them. The only problem now is that Micron have stopped selling E die for cheap or, at least they mix in lower performing chips in the standard Ballistix range as well as reserving the best of E die for ballistix max, so stuff like Team Group and G Skill are the only way to guarantee Samsung B die ... Or £470 for 2x16 Gb Ballistix Max ewwww. Micron gotten greedy.

Have you seen anyone that have killed their DDR4 ram yet from 1.5-1.55v? Lots of enthusiasts are doing just that with these two chips, as well as newer high end kits coming rated for 1.5v by default.

 

[Guest2]

I thought Micron E can take temps in excess of 70c at 1.5V? There was a vid of some Micron E being heated up with a hair dryer (at least I think it was Micron E) 70c and memtest was still running fine with no errors

Would be interested to know if Samsung B die can take much more than 1.5v

 

[Guest2]

I updated my bios today and now I can run FLCK 1900, I haven't tried any higher. I ran the same timings I previously ran when my 3700X worked at 1900, though I tuned back down then as it wouldn't always post.

How does this look? Anything else I should try tweaking?

Do you have you 5800X running PBO auto or manual OC (either manual all core or PBO enabled with anything else)
Interested to know whay read and write speeds are higher. Slightly higher latencies will be down to timings but read/write maybe the boost from CPU as im running 3800 / IF 1900 too

 

[Nick1881]

Do you have you 5800X running PBO auto or manual OC (either manual all core or PBO enabled with anything else)
Interested to know whay read and write speeds are higher. Slightly higher latencies will be down to timings but read/write maybe the boost from CPU as im running 3800 / IF 1900 too

I had PBO and everything auto for that, no kind of overclock. Now I have started playing with curve optimiser and running up to 5Ghz, I should try another run.

Edit - some speeds are slower now

 

[Guest2]

I had PBO and everything auto for that, no kind of overclock. Now I have started playing with curve optimiser and running up to 5Ghz, I should try another run.

Edit - some speeds are slower now

Check R20 scores and AIDA64 cache & memory benchmark with pbo on auto vs other settings.
AMD does something like appears to boost mhz values to 5Ghz + but its only numbers and nothing else.
There’s a name for it but forgot what it is ‘AMD mhz value only booster’ would be accurate

 

[GregI]

Would be interested to know if Samsung B die can take much more than 1.5v

I might be crazy but I remember a lot of 1.55v XMP from G. Skill. I don't see any of these anymore. I wonder if they noticed a larger amount of RMA for the 1.55v kits and decided not to go after that anymore. I read a thread ages ago where someone used crappy B-Die sticks at 1.6v. They died after a year and a half.

1.50v is still being made so we can assume it's fine. The issue is I don't know much about if different B-Die bins have different voltage and temperature tolerances. Would be weird for that to be a thing.

 

[Nick1881]

Check R20 scores and AIDA64 cache & memory benchmark with pbo on auto vs other settings.
AMD does something like appears to boost mhz values to 5Ghz + but its only numbers and nothing else.
There’s a name for it but forgot what it is ‘AMD mhz value only booster’ would be accurate

I ran R20 and single core was higher. Sadly after 5 minutes my system crashed during COD MP, so I've gone back to stock. I think I'm just going to leave it alone, it's already plenty fast enough. I run enough to keep my frames over 144 and that's all I need. It's fun tinkering though. I'm also happy my ram and IF are 3800/1900.

 

[Guest2]

I ran R20 and single core was higher. Sadly after 5 minutes my system crashed during COD MP, so I've gone back to stock. I think I'm just going to leave it alone, it's already plenty fast enough. I run enough to keep my frames over 144 and that's all I need. It's fun tinkering though. I'm also happy my ram and IF are 3800/1900.

I’ve done the same CPU wise, left it on PBO auto. Will try again with curve optimiser when the new BIOS are released that supports PBO2.
3800 / 1900 if CAS 14 has seemed to make a difference by itself so happy with that. Might try again with new BIOS for 4000 / 2000 if cas 16. I know this RAM can do 4000Mhz at least

 

[BallistixOnZ490]

I might be crazy but I remember a lot of 1.55v XMP from G. Skill. I don't see any of these anymore. I wonder if they noticed a larger amount of RMA for the 1.55v kits and decided not to go after that anymore. I read a thread ages ago where someone used crappy B-Die sticks at 1.6v. They died after a year and a half.

1.50v is still being made so we can assume it's fine. The issue is I don't know much about if different B-Die bins have different voltage and temperature tolerances. Would be weird for that to be a thing.

So 1.5v for Samsung B die, 1.55v for Micron E die seem to be well tolerated maximums. Samsung typically manages much better secondary timings, and Crucial are overcharging for binned 4000+ kits now.

Back with DDR3 loads of enthusiasts were frying Elipda chips trying to push 2400 Mhz CL9, but Im seeing any reported failures at the 1.5 - 1.55v range for these current DDR4 chips.

 

[Alyjen]

Check R20 scores and AIDA64 cache & memory benchmark with pbo on auto vs other settings.
AMD does something like appears to boost mhz values to 5Ghz + but its only numbers and nothing else.
There’s a name for it but forgot what it is ‘AMD mhz value only booster’ would be accurate

you think about core stretch? when it's reporting high clocks but effective clocks are way lower? you can check it via HWiNFO, just need to watch it during the load to see that effective clocks never reach what it was showed as boost clock

 

[Heofz]

i heard there were issues on MSI x570 boards in running memory faster than 3200mhz but not sure if that's been resolved since... I've currently got the 3200mhz C14 8pack 2x16 kit... worth upgrading to the 3600mhz C16 set?

 

[Joxeon]

i heard there were issues on MSI x570 boards in running memory faster than 3200mhz but not sure if that's been resolved since... I've currently got the 3200mhz C14 8pack 2x16 kit... worth upgrading to the 3600mhz C16 set?

The 3200/14 and 3600/16 are essentially the same kit but with a different XMP profile programmed in so you could just set your current kit to 3600/16.

 

[Guest2]

you think about core stretch? when it's reporting high clocks but effective clocks are way lower? you can check it via HWiNFO, just need to watch it during the load to see that effective clocks never reach what it was showed as boost clock

Yeh, that's it, thanks. core stretch indeed

i heard there were issues on MSI x570 boards in running memory faster than 3200mhz but not sure if that's been resolved since... I've currently got the 3200mhz C14 8pack 2x16 kit... worth upgrading to the 3600mhz C16 set?

I have an MSI x570 board, was fine at XMP (3600Mhz) 3800 and 4000

Cant you just try running the 3200Mhz kit at 3600 C16? Then push see if it pushes higher, like 3600 C14 then 3800 C14. I would have thought 3600 C16 kit is about the same spec as 3200 C14

 

[Heofz]

The 3200/14 and 3600/16 are essentially the same kit but with a different XMP profile programmed in so you could just set your current kit to 3600/16.

Cant you just try running the 3200Mhz kit at 3600 C16? Then push see if it pushes higher, like 3600 C14 then 3800 C14. I would have thought 3600 C16 kit is about the same spec as 3200 C14

I could... I just don't really trust myself using the DRAM calculator for all the other timings, and haven't looked into how the RF ratio thing works

 

[Joxeon]

I could... I just don't really trust myself using the DRAM calculator for all the other timings, and haven't looked into how the RF ratio thing works

Just put the main timings in from the 3600 mhz version so 16-16-16-16-38 and the board will handle the rest.

 

[bentnose555]

Anyone know how much of a difference i would see in games (mainly Cod & Warzone) with a 5900x paired with C16 2666mhz upgrading to the team group C16 3600mhz?

Will i see an increase in frames do you think? Trying to hit the magic 165hz

 

[GregI]

Aida64 updated: https://www.aida64.com/news/aida64-v632-amd-zen-3-optimizations-and-amd-radeon-rx-6000-series

Has some improvements for Zen 3.

"GPU details for nVIDIA GeForce RTX 3050, GeForce RTX 3060, GeForce RTX 3060 Ti, GeForce RTX 3070 Ti, GeForce RTX 3080 Ti" lol