New microcode fix for 13th/14th gen. Thoughts?

Gorbash12346 · 3 Oct 2024 at 23:16

Just wondering how everyone else is fairing with the new microcode (honest it's fixed this time)
I've had my replacement 13900K for a couple of weeks (took over 2 months to get replaced) and initially it was 'ok' occassionaly it would boost to 5.8ghz on 2x cores with temps maxing out around 78 degrees C but since applying this update it will not boost either of the preferred cores above 5.5 at any point and the remaining cores are dropping back below 5ghz when it does. Thats a significant performance loss is it not?

I'm also getting issues with running my audio interface for my guitar with significant audio drop out where I had none previously. I've had to resort to using my I7-10750H laptop for all guitar stuff which is really not ok.

Anyone else having similar issues? any suggestions?

G J · 4 Oct 2024 at 00:06

Is this is the latest 0x12B microcode update as I think its only released at the moment for B760/Z790 boards so I have to wait for the moment for my MSI Z690 A-Pro only has the 0x129 update so it may be a while for others to get the 0x12B update and test it.

Gorbash12346 · 4 Oct 2024 at 00:38

Yes. Thats the one.

Journey · 5 Oct 2024 at 09:59

Gorbash12346 said:
Anyone else having similar issues? any suggestions?

Naff all you can do. Your pretty much resigned to having to accept you paid for a level of performance you won't ever get back, and you won't get a refund either so not a lot of options sadly.

Clarence Boddicker · 11 Oct 2024 at 08:48

Ive got a 14700k and updated to the new microcode yesterday. No performance loss, but voltage is still going up to 1.45v with the intel default settings, the same as previous microcode. I thought that it was meant to address these high voltages?

Daytrader · 11 Oct 2024 at 09:16

I thought as long as it did not go over 1.55 all was ok, i go to like 1.49 sometimes.

VincentHanna · 15 Oct 2024 at 18:41

I just updated the BIOS on my Z690 motherboard for the first time in about 18 months (13700k)

I did a before and after quick CPU Z Benchmark and lost about 100 points on multicore but single core performance was the same.

I mainly game on it, so I don't expect to see any dramatic change.

Jay85 · 15 Oct 2024 at 19:09

VincentHanna said:
I just updated the BIOS on my Z690 motherboard for the first time in about 18 months (13700k)

I did a before and after quick CPU Z Benchmark and lost about 100 points on multicore but single core performance was the same.

I mainly game on it, so I don't expect to see any dramatic change.

Did the same but for overclocking reasons to see if it added more stability for memory OC. So far no change, OC is still stable at least so that's one thing less to worry about. Didn't see any change in performance personally either.

dholdi · 15 Oct 2024 at 19:44

Is this done via a bios update ?

Edit - just checked mb website and it seems so, theres ea few available for mine now.

Vimes · 16 Oct 2024 at 10:51

I noted that Gigabyte did another variant of the F30 BIOS, including 0x12B microcode, for my Z690 board. That would be using a 14700k.

With it applied and Intel defaults used.....

Without Intel defaults....

Without defaults but 253W power limits applied...

Only one measure, not sure how much it would mean when in use day to day etc, I suppose type of use could have some influence.

This RL platform has been most.......interesting...!

Rroff · 16 Oct 2024 at 12:15

No real difference to performance but mine seems to be running slightly higher volts and slightly higher wattage with the update than before - with 286 watt limit set before it would hit about 284 in HWMonitor now seeing up to 312.

Vimes · 16 Oct 2024 at 12:31

Rroff said:
No real difference to performance but mine seems to be running slightly higher volts and slightly higher wattage with the update than before - with 286 watt limit set before it would hit about 284 in HWMonitor now seeing up to 312.

Is that 312w maximum so far noticed in something like CBr23 with 286w profile set, as the maximum for the two power stages..? If so would that be expected, to see up to 312w..?
I note that if I set mine to 253w then that is it, HWMonitor does not show anything above 253w.

Rroff · 16 Oct 2024 at 12:38

Vimes said:
Is that 312w maximum so far noticed in something like CBr23 with 286w profile set, as the maximum for the two power stages..? If so would that be expected, to see up to 312w..?
I note that if I set mine to 253w then that is it, HWMonitor does not show anything above 253w.

If I run CB23 it will peak at 312 though most of the time it doesn't go over 286, before it would peak at like 284.xx. I've not played around much to see what is going on though.

Vimes · 16 Oct 2024 at 12:53

Rroff said:
If I run CB23 it will peak at 312 though most of the time it doesn't go over 286, before it would peak at like 284.xx. I've not played around much to see what is going on though.

Is that with no maximum limits set..?

It did surprise me to note in one video I watched this morning it noted 442 total power consumption when running R23 with a 14700k....

I'm not overly focussed, but curious enough perhaps to appreciate there seems to be more going on with the Intel defaults to give such differences in in the CB screens I posted than just that of power limits.
Then again I do not know how each board manufacturer abide by or implement what Intel have decided are their suggested / recommended default values.

Rroff · 16 Oct 2024 at 12:55

Vimes said:
Is that with no maximum limits set..?

Supposedly 286 watt PL2 but it seems to be acting kind of like it is set to 320, not sure if Gigabyte has done something wrong or whether I need to mess about with settings a bit.

EDIT: XTU is showing Short Power Max as unlimited so something isn't working right.

Vimes · 16 Oct 2024 at 13:42

Rroff said:
Supposedly 286 watt PL2 but it seems to be acting kind of like it is set to 320, not sure if Gigabyte has done something wrong or whether I need to mess about with settings a bit.

EDIT: XTU is showing Short Power Max as unlimited so something isn't working right.

Whether needed or appropriate I set both PL1 and 2 to 253w within the Gigabyte BIOS and that seems to be applied and held.

Rroff · 19 Oct 2024 at 13:54

Vimes said:
Whether needed or appropriate I set both PL1 and 2 to 253w within the Gigabyte BIOS and that seems to be applied and held.

Was operating as if unlimited, not sure what was going on exactly as the settings in the BIOS and XTU and actual behaviour were different, but due to air cooling was effectively being limited at around 310-320 watt (on water would have probably gone 400+ like screenshot above), I've just enabled Intel's performance profile for now with 253w limit until I've got time to dabble a bit - Gigabyte's optimised setting with 286 watt limit was working well for me before and should still avoid any degradation issues *apparently*.

Vimes · 19 Oct 2024 at 15:03

Rroff said:
Was operating as if unlimited, not sure what was going on exactly as the settings in the BIOS and XTU and actual behaviour were different, but due to air cooling was effectively being limited at around 310-320 watt (on water would have probably gone 400+ like screenshot above), I've just enabled Intel's performance profile for now with 253w limit until I've got time to dabble a bit - Gigabyte's optimised setting with 286 watt limit was working well for me before and should still avoid any degradation issues *apparently*.

I am not overly familiar with XTU, but could that have made changes to what the default Intel profile was when selected..? Assuming that you have checked that tho.
Running R23 with the 253W max settings, as noted, will show the 253w as the maximum attained, within HWiFO64, as well as being the reason for the limitation.

I am unsure, some doubt was cast when @Grim5 posted a BZ video, not with the latest microcode, to suggest that Intel's Default profile is needed for the microcode to work....

14th Gen "Raptor Lake Refresh"

Have you taken a performance hit by doing so? I haven't benched to notice. But more than likely I have.

forums.overclockers.co.uk

not sure if that has changed.

The whole situation is messy.

jigger · 19 Oct 2024 at 16:03

Intels boost algorithm is opportunistic and that’s baked into the chip. Intels fix is to gain some continuity between the motherboard manufacturers power profiles. The chip itself remains opportunistic.

Clarence Boddicker · 19 Oct 2024 at 16:07

Vimes said:
The whole situation is messy.

It is indeed. I’m getting lower voltages with the intel default settings disabled, and gigabytes perfdrive optimisation selected. I’ve got no idea if it’s the right thing to do for the longevity of the cpu though. Ill stick with the intel settings disabled for now, surely lower voltages are safer.