Jump to content
Search In
  • More options...
Find results that contain...
Find results in...


  • Content Count

  • Joined

  • Last visited


This user doesn't have any awards

About Rym

  • Title

Recent Profile Visitors

The recent visitors block is disabled and is not being shown to other users.

  1. RTX 3060 units which have been used for mining are now being sold for 270$ a piece on the chinese second hand market. "The Block is reporting that a "large batch" of GeForce RTX 3060 graphics cards with prices ranging anywhere between $336-$450 or so... but the seller didn't get many buyers, so the prices were reduced to $270. Another seller is getting rid of every single gaming laptop they purchased for crypto mining, and selling them for around $1000" Source: https://www.tweaktown.com/news/80492/crypto-crackdown-geforce-rtx-3060-now-costs-270-second-hand-in-china/ind
  2. I'm not sure what you said tbf. I know curve optimizer helps in higher end chips, but all my tests with it either proved unstable or gave no performance benefit. With the auto overclock feature in PBO maxed out at +200MHz I get a stable single core boost of 4850MHz and an all-core boost of 4.6-4.65GHz in stress tests. I wanted to know if it's possible to override the maximum limit of +200MHz on the Auto OC to, say +250 or +300, +350MHz, which would likely enable me to reach up to 5GHz on single core because of massive thermal headroom and a substantial power
  3. I have a 5600X boosting to 4850MHz at 50~ degrees, and ClockTuner by 1usmus noted it could go up to 5050MHz. While CTR overclocking is unstable at best, it did run WoW Classic at 4900GHz with CTR undervolting it to 1.375 as well, but the undervolting it does causes hard reboots in benchmarks (cinebench). Since I know the CPU is capable of boosting past 4850MHz, is there any way to bypass the hard-cap AMD put up without resorting to CTR?
  4. I see. Well i don't mind considering there is zero difference between the 5600x and the 5950x in gaming outside of the difference caused by clock speeds. I think AMD did a good job here for gaming performance.
  5. When it really doesn't? For reference, I'm using a ryzen 5 5600X, I have PBO limits fine-tuned and auto overclock at +200Mhz. The CPU reaches 60 degrees in full load, 50~ in single core and boosts to 4.85GHz single core, 4.6 multicore stress testing Cinebench R23. I saw people on reddit praising curve optimizer saying it increases performance and how great it is. Why? It doesn't increase any performance for me, I never saw an undervolt increase performance, but it increased instability. I turned it off and stress test performance is the same as with curve turned on.
  6. You look like you have a lot of stuff in your build. A system restart will fully refresh your system (a shutdown does NOT do this). It is unclear whether your problem is caused by a program or hardware power saving mode or otherwise limitation. If it's not too much bother, I'd just do a quick system restart every day if you NEED to keep it active day and night.
  7. Okay so the direct physical contact of the mounting bracket won't cause issues?
  8. Without installing the pump itself it's already less than half a mm away from direct physical contact.
  9. Hey, long story short: The Arctic Liquid Freezer II 360 Revision 4 cooler I ordered arrived yesterday, I tried installing it and after following the guide at https://support.arctic.de/lf2-360r4 as well as https://www.youtube.com/watch?v=ahbzLFUqLnk I noticed that my mounting bracket is extremely close to the capacitors near my VRMs, with about 1mm or less distance between the two. Now the mounting thingy is made out of metal, when I noticed how close they were I aborted the installation as I was unsure if it would cause it to short or not. It looked like so: https://cdn.d
  10. IQ tests do not display your real intelligence, and are incapable of displaying your EQ which is more important.
  11. Zen 5 Raphael will be another massive improvement in performance and efficiency. You futureproof your system by reselling your current gen CPU and mobo so you can buy the next generation.
  12. So I upgraded to a Ryzen 5 5600X, and Ryzen Master says cores 02 and 04 are my best cores, 04 being gold star and 02 being silver. But these very same cores are the ones that support the LEAST amount of negative curve optimizer offset, every other core goes to -20 but these 2 best cores can cause Cinebench R23 to crash even at -5 offset. Aren't these cores supposed to support the MOST negative optimizer offset?
  13. Not all systems suffer from tearing. Mine rarely does. Not all games cause tearing either. Gsync is good for single player story games where you can also pull out the 144 frames per second. Obviously this doesn't really happen in new games, even with an RTX 3090. G-sync used to be a marketing hyper back in the day, but it's just Vsync with a few added quirks. It may make your experience smoother but in my experience lower monitor refresh rate = stuttery mouse and image, so that a 50Hz refresh rate 50 FPS is worse than 144Hz refresh rate 50 FPS.
  14. So i should run them in a 16GB + 8GB combination instead of 16+16 on one channel and 8+8 on the other? Well it does make sense actually.
  15. Long story short i bought 2 16GB crucial ballistix ram modules, each at 3200mhz CL16 but they're single rank, so i want to fill out the other 2 ram slots with 2 single ranks modules to get dual rank benefits on my ryzen CPU. I could buy a kit of 2x 8GB crucial ballistix at identical speeds and latency. I suspect they will work together from the experience of the 3 ram modules i have used so far, but I'd like further opinions. Should i expect the setup to function normally as a 4 sticks dual rank setup, where the only major difference between them is their si