I9 9900k

Nickolp1974 · 3 Dec 2018 at 16:04

just did 2 quick tests, 1st with low mem/cache and 2nd higher on both and these are both the 1st runs after booting. there are gains to be had but not worth much...

moorhen2 · 3 Dec 2018 at 17:01

A couple of quick runs using roughly your settings, getting roughly the same results, just using less vcore. My insider build of W10 was being a PITA, so booted into my normal Build.

Deleted User 456458 · 3 Dec 2018 at 17:11

@Nickolp1974 Why would my 5ghz be so close to your 5.2ghz score? You mentioned chip efficiency. Can you elaborate on that a bit?

I wonder what other CPU based tests I can run to see what's going on. Pifast maybe?

Nickolp1974 · 3 Dec 2018 at 17:28

Robert896r1 said:
@Nickolp1974 Why would my 5ghz be so close to your 5.2ghz score? You mentioned chip efficiency. Can you elaborate on that a bit?

I wonder what other CPU based tests I can run to see what's going on. Pifast maybe?

There are always lots of factors when running synthetic benchmarks, too name examples

1. Operating system, certain benchmarks prefer a particular OS, geekbench win 7, xtu win 10 32bit
2. Certain versions of a benchmark can make a huge difference
3. VGA used, older vga with low memory can help.
4. Screen resolution
5. Cinebench R15 for example, gets a better result with the window ran off screen

And many other variables, are you running win 8 Rob?

Deleted User 456458 · 3 Dec 2018 at 17:32

Nickolp1974 said:
There are always lots of factors when running synthetic benchmarks, too name examples

1. Operating system, certain benchmarks prefer a particular OS, geekbench win 7, xtu win 10 32bit
2. Certain versions of a benchmark can make a huge difference
3. VGA used, older vga with low memory can help.
4. Screen resolution
5. Cinebench R15 for example, gets a better result with the window ran off screen

And many other variables, are you running win 8 Rob?

Win10 latest updates with a 2080ti. Pretty standard. Jealous of your mem OC. My Aorus pro sucks at RAM tweaking and buildzoid confirmed this also as he has the same board

moorhen2 · 3 Dec 2018 at 17:49

Robert896r1 said:
@Nickolp1974 Why would my 5ghz be so close to your 5.2ghz score? You mentioned chip efficiency. Can you elaborate on that a bit?

I wonder what other CPU based tests I can run to see what's going on. Pifast maybe?

Give wPrime a go.

moorhen2 · 3 Dec 2018 at 18:12

Nickolp1974 said:
Great volts for 5.2 but whys it scoring like it's at 5ghz?.? CPU looks great, show your max poss score, gotta be 2400+ on normal water

2400+ at 5.2 is wishful thinking I would say. Even with the cache at 50 only nets 2300 in Cinebench.

Nickolp1974 · 3 Dec 2018 at 18:14

moorhen2 said:
Give wPrime a go.

Yeah 1024, or one of the oldest and still the best is superpi32m, this will show how good your efficiency is, mems have to be tuned to perfection and be rock solid. If you didnt know Windows XP is always gonna be best but tbh who uses that apart from hard-core tweaker!
As most here are on win 10, let's have a go but cap clocks on CPU to say 50/47?? Anyone wanna play?

Nickolp1974 · 3 Dec 2018 at 18:16

moorhen2 said:
2400+ at 5.2 is wishful thinking I would say. Even with the cache at 50 only nets 2300 in Cinebench.

No, I meant judging by your cpu which has nice low volts a max poss score should be 2400+ pts@ around 5.4ghz

moorhen2 · 3 Dec 2018 at 18:26

Nickolp1974 said:
No, I meant judging by your cpu which has nice low volts a max poss score should be 2400+ pts@ around 5.4ghz

My mistake, crossed wires again.

Deleted User 456458 · 3 Dec 2018 at 18:26

Nickolp1974 said:
Yeah 1024, or one of the oldest and still the best is superpi32m, this will show how good your efficiency is, mems have to be tuned to perfection and be rock solid. If you didnt know Windows XP is always gonna be best but tbh who uses that apart from hard-core tweaker!
As most here are on win 10, let's have a go but cap clocks on CPU to say 50/47?? Anyone wanna play?

Don a run and screenshot the parameters. I’m in.

Nickolp1974 · 3 Dec 2018 at 18:34

moorhen2 said:
My mistake, crossed wires again.

heres a quick example i have just done no tweaks or anything

moorhen2 · 3 Dec 2018 at 19:35

Nickolp1974 · 3 Dec 2018 at 19:41

moorhen2 said:

Good, any OS tweaks yet?? Diagnostic mode, maxmem 4096gb, priority realtime?

moorhen2 · 3 Dec 2018 at 19:53

Nickolp1974 said:
Good, any OS tweaks yet?? Diagnostic mode, maxmem 4096gb, priority realtime?

No tweaks of any kind, we need to keep a level playing field, no tweaking tersiary's or RTL' etc. or sub-timings.

Nickolp1974 · 3 Dec 2018 at 20:05

moorhen2 said:
No tweaks of any kind, we need to keep a level playing field, no tweaking tersiary's or RTL' etc. or sub-timings.

Yep ok, now keep cpu at 50/47 and tweak mems/OS to improve time, as long as core is at 5.0ghz and cache at 4.7ghz and Windows 10 is used then anything goes.

Nickolp1974 · 3 Dec 2018 at 20:08

for instance this is done in diagnostic mode and priority realtime, took the last 45 mins to pass got there in the end, next stop 4133mhz C12

Deleted User 456458 · 3 Dec 2018 at 21:38

Stock windows settings. No optimization of any sort. @Nickolp1974 can you share your RAM timings and volts for CL14 @ 3600? I will try. I have a bdie kit.

Nickolp1974 · 3 Dec 2018 at 22:11

All I set with those is primary and trfc so 14 14 34 2t 360 all the rest on auto
Vccio and SA at 1.2v, dram 1.5v

That's all not tweakeď or anything and booted straight in, the Gene is very good with its memory.

moorhen2 · 4 Dec 2018 at 09:46

Nickolp1974 said:
All I set with those is primary and trfc so 14 14 34 2t 360 all the rest on auto
Vccio and SA at 1.2v, dram 1.5v

That's all not tweakeď or anything and booted straight in, the Gene is very good with its memory.

Then you are both using 3600mhz ram, I am using 3200mhz ram, so your overclocking on the ram from 3600 to 4000 is very good. But my 3200 is matching your scores at 3600.