AMD Ryzen 7 5800X 3D Cache Eight Core 4.5GHz (Socket AM4) Processor - Retail - Go Go Go xD

[Guest2]

Going on this GN video it circumvents some of AMD's built-in protections.

Level 3 manages to all core boost with less voltage and an 11c drop in temps. I'll stick with that for now.
68.5c max playing a game or Warzone 2

@Woodsta888 Thats the thing though, my fps was capped at 97fps on the 5800X chip and is on the 5800X3D. During gaming when it doesnt change from 97fps, surely the 1% lows do not come into play (because its 97 all the time)
...yet the 5800X3D feel smoother. I dont think its realted to 1% lows, unless in game fps counter isnt reporting 1% lows

 

[Guest2]

Looks like the kombo option is similar to undervolting?

I expect it is, just built into the BIOS so is better IMO

 

[Nexus18]

I expect it is, just built into the BIOS so is better IMO

Yup would have given it a go myself but got pbotuner setting to -30 on all cores so can't be bothered testing if it is the same thing as undervolting

 

[Guest2]

Yup would have given it a go myself but got pbotuner setting to -30 on all cores so can't be bothered testing if it is the same thing as undervolting

Do you run PBOTuner 2 every time you load windows or use Powershell / Scheduled task with PS script to start it?
Also, Kombo strike is for MSI motherboards only (in case you are something other than MSI)

If you were MSI and messed with PBO tuner every time loading windows I and a single option did the same, I would go with the single option personally (maybe save test results from PBO tuner and compare to Kombo level 3)

 

[NightSt@lk3r]

Do you run PBOTuner 2 every time you load windows or use Powershell / Scheduled task with PS script to start it?
Also, Kombo strike is for MSI motherboards only (in case you are something other than MSI)

If you were MSI and messed with PBO tuner every time loading windows I and a single option did the same, I would go with the single option personally (maybe save test results from PBO tuner and compare to Kombo level 3)

I use scheduled task to apply it every boot.

 

[Nexus18]

Do you run PBOTuner 2 every time you load windows or use Powershell / Scheduled task with PS script to start it?
Also, Kombo strike is for MSI motherboards only (in case you are something other than MSI)

If you were MSI and messed with PBO tuner every time loading windows I and a single option did the same, I would go with the single option personally (maybe save test results from PBO tuner and compare to Kombo level 3)

As @NightSt@lk3r said. Works well.

Have a MSI MB too and did notice that kompu feature, thought it might have been some auto OC/boost kind of feature. Having a look at that reddit thread, it looks pretty similar to a -30 undervolt i.e. lower temps and power.

 

[varkanoid]

Hhhhmm I'm still hitting low 90s even with load optimized defaults in the BIOS and using a PBO curve, any ideas what else it could be? Perhaps reseating my HSF, but it seems to be behaving in other situations (gaming, idles etc).

You need to make sure your cooler is optimised for these CPU's. Not sure what you have but it can make difference of 8-10c. Just as an example here is a bit from my AIO instructions.
I've read some people just drill the extra holes themselves.

 

[doodah]

Thanks - this is with GTA V running the benchmark @ 1440p, CS Go was in the 50-60C range I think.

 

[Guest2]

Cyber Monday has been disappointing for this chip.
Suppose it’s all down to supply and demand

 

[Journey]

Cyber Monday has been disappointing for this chip.
Suppose it’s all down to supply and demand

You are a week early buddy.

 

[true_gamer]

Thought I would share my dialed in 24/7 stable ram timings/voltages with no WHEA errors. Worked well with my old 5800X and the new 5800X3D. Seems to be the sweetspot without using higher voltages.
Make sure to pay attention to the VSOC, VDDP, VDDG voltages, as these had to be manually set to be stable.
3733 14 13 13 13 28 42 256 1T 1.45V
(Patriot Viper 4400Mhz 2x8GB kit)

You can use this leaderboard spreadsheet to add your timings etc. https://docs.google.com/spreadsheet...0MWVPcYjf6nOlr9CtkkfN78tSo/edit#gid=197347422



Here's my 4000 16 16 16 16 32 48 288 1T 1.45V - Although this was stable with no WHEA errors on my system, I wouldn't advise running these timings 24/7.



For those wondering if upgrading from a 5800X to a 5800X3D is worth it at 4K? Absolutly!

Not a fair comparision with the swap of GPU's from a 3080 Ti to a 4090, but when comparing the CPU Game/CPU Render, theres a huge uplift over the other even at 4K.

 

[Guest2]

You are a week early buddy.

Are, thought it was the Monday before black friday. oops

 

[michty_me]

Thought I would share my dialed in 24/7 stable ram timings/voltages with no WHEA errors. Worked well with my old 5800X and the new 5800X3D. Seems to be the sweetspot without using higher voltages.
Make sure to pay attention to the VSOC, VDDP, VDDG voltages, as these had to be manually set to be stable.
3733 14 13 13 13 28 42 256 1T 1.45V
(Patriot Viper 4400Mhz 2x8GB kit)

You can use this leaderboard spreadsheet to add your timings etc. https://docs.google.com/spreadsheet...0MWVPcYjf6nOlr9CtkkfN78tSo/edit#gid=197347422



Here's my 4000 16 16 16 16 32 48 288 1T 1.45V - Although this was stable with no WHEA errors on my system, I wouldn't advise running these timings 24/7.



For those wondering if upgrading from a 5800X to a 5800X3D is worth it at 4K? Absolutly!

Not a fair comparision with the swap of GPU's from a 3080 Ti to a 4090, but when comparing the CPU Game/CPU Render, theres a huge uplift over the other even at 4K.

Some good results there! Does having really tight timings on the X3D give much of a performance gain? I was always under the impression that it didn't really matter much?

Edit: I was one of the ones still sitting on the fence on this but for 1440p.

 

[Guest2]

Some good results there! Does having really tight timings on the X3D give much of a performance gain? I was always under the impression that it didn't really matter much?

Edit: I was one of the ones still sitting on the fence on this but for 1440p.

Im interested in that answer too, I expect very little gains but could be wrong. If I am ill tweak my RAM

 

[true_gamer]

OK, so I have ran 9 test... To see if Ram Clocks/Timings make a difference with the 5800X3D.

5800X3D -30 CO 4450MHz all core boost.
4090 FE Stock.

4K 3733 14 13 13 13 28 42 1T


4K 3600 16 16 16 32 48 1T


4K 3200 14 13 13 13 28 42 1T

 

[true_gamer]

1440p

1440p 3733 14 13 13 13 28 42 1T


1440p 3600 16 16 16 32 48 1T


1440p 3200 14 13 13 13 28 42 1T

 

[true_gamer]

1080p

1080p 3733 14 13 13 13 28 42 1T


1080p 3600 16 16 16 32 48 1T


1080p 3200 14 13 13 13 28 42 1T

 

[michty_me]

1080p

1080p 3733 14 13 13 13 28 42 1T


1080p 3600 16 16 16 32 48 1T


1080p 3200 14 13 13 13 28 42 1T

That's brilliant, thanks for carrying out that tests. So pretty much margin of error differences between them all!

 

[Guest2]

That's brilliant, thanks for carrying out that tests. So pretty much margin of error differences between them all!

Agreed, great tests @true_gamer
Thanks.

31fps higher cpu high on 1080p 3600c16 than 3733c14
Margin of error still?

I read Shadow of the Tomb Raider also favours lower latency memory.

Based on this, i’ll probably leave my RAM as it is at XMP (3600C16/1800IF) or manually set it to 16 16 16 32 48 and turn off GDM

 

[michty_me]

Agreed, great tests @true_gamer
Thanks.

31fps higher cpu high on 1080p 3600c16 than 3733c14
Margin of error still?

Good catch, never noticed that actually. In my defense it was the middle of the night haha.