[SOLVED] XMP and heat stability

[Assaf Patishi]

Hi everyone,
I only recently learned about heat stability. I would like to know if this is an issue if you are using the XMP profile.
All the ram companies (G.skill, Crucial etc'), how can they guarantee stability if they don't know what is the heat situation in one's pc? (what case you use, if you have enough airflow...). The QVL states the motherboards which has been tested to work properly with certain kits, but they don't say anything about temperature.
I also noticed that some timings are left at JEDEC specs' like trfc (500,600,700..depending on the frequency) and command rate (usually at 2) which lead me to think that they leave this on purpose to make the ram be more stable (for heat?).
Are those timings really affect heat stability? what other settings can affect this?

Thx.

 

[Eximo]

Been a while since memory temperatures had much of an effect. Early DDR3 could run a little warm, but a few years in the process nodes were quite efficient. Memory heat spreaders are more for looks than anything else these days. DDR4 and DDR5 run at such relatively low voltages that they don't generate enough heat that their surface area can't dissipate.

It would take quite the hot box to cause issues with memory temperatures.

Timings vary a lot between sticks, some are tuned for stability with high timings. Others are tuned for extreme speeds. They all get manufactured, tested, sorted before being sold, so really the timings come down to luck. Now they will target the most profitable market segments, so sometimes better memory sneaks into lower grade batches. Not much point making DDR4-4000 kits if you only get 1% the sales as 3600. So the best of the best get put in the premium products, and left overs filter down into the lower tiers.

 

[Assaf Patishi]

Thank you for your answer.
I actually heard that Samsung B-die chips are very sensitive to heat and that they can put out errors at 50+c (which otherwise be completely stable). My 4000 CL16 kit is also Samsung B-die but I run it at XMP settings and I never had issued even at 50+c.
I was just wondering if this can be a result of the high trfc and command rate.
A lot of overclockers usually set trfc to around 300, and XMP profile default is 700 (for 4000mhz). Also many overclockers try to run the memory at command rate 1 (instead of 2). So I was wondering maybe this is part of the reason for getting the errors.