F5F Stay Refreshed Hardware Desktop i9 9900k performance tips Optimize settings for better speed and efficiency.

i9 9900k performance tips Optimize settings for better speed and efficiency.

i9 9900k performance tips Optimize settings for better speed and efficiency.

Pages (3): Previous 1 2 3
L
louisryder
Junior Member
45
04-22-2018, 03:56 PM
#21
I think this video will be interesting to watch. Thanks. It seems a bit off, but the bigger kits are easy to spot just by checking the timing charts, which is exactly what I need. I linked AMD random restarts to Hynix memory modules, especially the lower-end ones. The 3000mhz LPX Vengeance model works well—I have two sets. At default settings, it shows rare random restarts. With XMP, random restarts happen more often. After replacing them with 3600mhz CL16-16-16-36 B-Dies, I didn’t notice any problems. Tested across four or five processors: 1200, 1400, 200ge and 220ge. I used them with 2700x, but performance was terrible regardless. I haven’t used Micron at all for DDR4 yet. The E-dies from these seem better due to improved scaling. Smaller timings aren’t the only factor—many people overlook how CPU overclocking affects memory latency.
L
louisryder
04-22-2018, 03:56 PM #21

I think this video will be interesting to watch. Thanks. It seems a bit off, but the bigger kits are easy to spot just by checking the timing charts, which is exactly what I need. I linked AMD random restarts to Hynix memory modules, especially the lower-end ones. The 3000mhz LPX Vengeance model works well—I have two sets. At default settings, it shows rare random restarts. With XMP, random restarts happen more often. After replacing them with 3600mhz CL16-16-16-36 B-Dies, I didn’t notice any problems. Tested across four or five processors: 1200, 1400, 200ge and 220ge. I used them with 2700x, but performance was terrible regardless. I haven’t used Micron at all for DDR4 yet. The E-dies from these seem better due to improved scaling. Smaller timings aren’t the only factor—many people overlook how CPU overclocking affects memory latency.

S
SuperMarioDXB
Member
243
04-22-2018, 05:05 PM
#22
Attempted 48 cache setting without modifying others, ambient high maintained.
S
SuperMarioDXB
04-22-2018, 05:05 PM #22

Attempted 48 cache setting without modifying others, ambient high maintained.

J
jan30els
Junior Member
5
04-23-2018, 01:44 AM
#23
Thanks for the info. I attempted Prime95 with 46cache on 1.28v and it froze, then switched to 47cache on 1.3v which also caused issues. At 5.1ghz with 45cache on 1.31v, it ran smoothly for an hour without any problems. It seems challenging to boost cache without significantly raising the voltage, unlike just adjusting core clocks. I haven’t checked how much cache is needed yet.
J
jan30els
04-23-2018, 01:44 AM #23

Thanks for the info. I attempted Prime95 with 46cache on 1.28v and it froze, then switched to 47cache on 1.3v which also caused issues. At 5.1ghz with 45cache on 1.31v, it ran smoothly for an hour without any problems. It seems challenging to boost cache without significantly raising the voltage, unlike just adjusting core clocks. I haven’t checked how much cache is needed yet.

M
MegaEdu32
Junior Member
1
04-23-2018, 03:35 AM
#24
managed to achieve 5.1ghz and 1.31v, though it appears challenging to increase the cache beyond 45
M
MegaEdu32
04-23-2018, 03:35 AM #24

managed to achieve 5.1ghz and 1.31v, though it appears challenging to increase the cache beyond 45

J
Jarzzermann
Posting Freak
788
04-23-2018, 10:13 PM
#25
When discussing voltages, are you referring to the BIOS settings, vdroop results, or the VID? For my 9900K running at 5GHz all-core (up to 5.3Ghz single), I set the BIOS voltage around 1.37V, with a modest load line calibration that lets it drop to approximately 1.26V in Prime 95 small FFT. The VID sits near 1.42V under those conditions. Running my cache at 4.8Ghz makes it seem more capable than yours, but I experience temperatures of 80-90°C per core while using liquid metal and a custom loop—so your cores appear hotter despite similar performance. In short, the benefits you see will be minimal, so I’m doing this mainly for curiosity. I’m using an Asus Maximus XI Hero, which is quite similar to your Z390-F model.
J
Jarzzermann
04-23-2018, 10:13 PM #25

When discussing voltages, are you referring to the BIOS settings, vdroop results, or the VID? For my 9900K running at 5GHz all-core (up to 5.3Ghz single), I set the BIOS voltage around 1.37V, with a modest load line calibration that lets it drop to approximately 1.26V in Prime 95 small FFT. The VID sits near 1.42V under those conditions. Running my cache at 4.8Ghz makes it seem more capable than yours, but I experience temperatures of 80-90°C per core while using liquid metal and a custom loop—so your cores appear hotter despite similar performance. In short, the benefits you see will be minimal, so I’m doing this mainly for curiosity. I’m using an Asus Maximus XI Hero, which is quite similar to your Z390-F model.

Pages (3): Previous 1 2 3