Asrock X370 Taichi Overclocking Thread - Page 739 - Overclock.net - An Overclocking Community
Forum Jump: 

Asrock X370 Taichi Overclocking Thread

Reply
 
Thread Tools
post #7381 of 7419 (permalink) Old 03-26-2020, 12:08 PM
Noctua Fanboy
 
polkfan's Avatar
 
Join Date: Apr 2015
Posts: 1,043
Rep: 16 (Unique: 13)
I don't see 3900X throttling in MT with some off-set thomasck MT in the same on my setup but the effective clock is up and finally once again hitting turbo more then once every 3 days on one core lol not even lying it never hit turbo even on that boostester it was on random extremely light usage.

3700X PBO (IF, 1900mhz)16GB G.SKILL TridentZ Series Samsung B-die 3800mhz 14-17-10-32-42
MSI GTX 1080 Gaming X
EVGA SuperNOVA 750 G3
Fractal Design R6(All fans replaced with 5 Noctua A14 fans)
H150i(Fans replaced with Noctua NF-A12)

Noctua fanboy detected! Fans all run at 500RPM
polkfan is offline  
Sponsored Links
Advertisement
 
post #7382 of 7419 (permalink) Old 03-26-2020, 02:20 PM
Stock is too casual~
 
Veii's Avatar
 
Join Date: Aug 2018
Location: Austria / Bulgaria
Posts: 719
Rep: 37 (Unique: 27)
Quote:
Quote:
What you might lack of, is enabling UnCore OC flag in AMD Overclocking
All depends where you changed SOC
with uncore oc enabled it seems to boots a bit more, but when there's nothing in the background, FireFox is a boost killer as far as I can notice here
UncoreOC is a flag to enable different voltages without forced bounding and auto scaling
Enabling it will allow you to change each voltage individually
Not having it enabled, means the cpu will do all by itself, and if you don't get the ranges right by yourself it will ignore it and use their
i wrote it couple of pages back while explaining how to track ACPI CPPC core rating in % , tracking good cores bad cores for voltage testing


Quote:
Quote:
"You can see auto scaling is on, because 1.1 VDDG and VDDP was requested, which means SOC had to be 1.2v"
auto scaling is caused by uncore oc? that was off all the time
Yes, not having uncoreOC mode, let's the board predict all the values
You get them wrong, it doesnt care ^^
But because you have your profiles and this strange bugs, polkfan doesn't have
I think you made a back and forth flashing bug

Quote:
Quote:
"Important is fixing procODT and CAD_BUS values, else you mess everything else up"
I'm using the same as before, dram calc values, and ram got over two hours of kahru mem test no problems
yea no , that doesn't work, RTT is kits exclusive, procODT and CAD_BUS ranges changed because memory training (PMU) got an update too
optimal procODT is now starting at 28ohm , with 30ohm as option, up to 34ohm for single ranked, or 48-60 ohm for dual ranked kits

Quote:
Quote:
"About realworld performance, we need SiSoftware Sandra Multi-Core Efficiency (detailed) reports, to see where the problem lies
Two things can and surely will happen
- There can be a 1:1 desync even when it is enforced as enabled
(noticable in 100ns Inter-Core Latency from furthest core inside one CCX to the furthest core inside another CCD inside the last CCX)
- you can actually be voltage and resistance throttled by the FIT module, if your procODT defaults to 60ohm or CAD_BUS is messed up
Also because CAD_BUS has new values, your memory OC can be unstable, because PMU got an update
well fully everything got an update"
gonna check sissandra soon. new procodt and cadbus? where to get info about that?
Only through me , or by reading a lot of 1usmus's findings
We are in a situation of a PreRelease AGESA update with literally everything changed
There is no guide, but threadripper was us always 1 step in-front, and now it begins to scale similar and behave similar inside CCX

Quote:
Quote:
If you are throttling at 75c already, it can be an SMU update
I haven't fully explored all the changes to this exact moment, too many things changed
But AMD CBS has i think under NBIO was it, a flag to specify new throttle temp
You can use 85c there
the only place I saw a flat to change throttling temperature is under pbo>advanced, nowhere else. I've looked around and found nothing, only a TDP but that would be in watts, not celsius
Yea no it's a real throttle temp location, it has it's own menu in amd CBS, you need to look closer , still think it was in NBIO next to max TDP rating in AMD CBS -> NBIO, somewhere there was it & CBS didn't change

Just yes, we are working here with cutting edge stuff, that isn't out or shouldn't be known that it's out
JZ was quiet on any information on it - and AMIBCP only shows "oh he used an offset, that's it"
While crosschecking this bios with the latest ones
a huge chunk of code was updated, starting to actually decompile it revealed all the changes
Sorry, there is no guide for something that shouldn't be out yet
But we already got to know how 1004B has to behave with resistance

I still need you guys to make reports with the SMU tool , so we check what's different
and it will be easier to 100% confirm an SMU version change
Or also confirm that there is a big change but it's hidden under the same version number
Anyways, SiSandra showed it clearly that fabric doesn't behave how it used to be and one limiter is gone

@polkfan
Can you reopen SiSandra, without doing a test - and just mark down the one testing of auto values 1900/3800 with 100ns inter-CCX latency
and the same one after you fixed voltages and it behaves how it should
I think , because you likely didn't change your setup, you should be able to use a very old result pre-tested vs the current record holding one with 66ns inter-ccx latency and 43ns intercore latency
just so we get a graph comparison , not only a text comparison
~ hmm i wonder if SiSandra on local score filter, allows us to import these exports as visual overview (detailed overview)

SiSandra Inter-Core Bandwith 14nm Ryzen global #1 | Twitter
Personal Records
RAM > Viper Steel @ 3466C14 (64.2ns) & Alt | Hynix-MFR 3333CL14 (72.5ns) & Alt

GPU > Gigabyte 970 G1 @ 1595|1900 (Hynix) | RX 580 Red Devil GS @ 1620|2200 (Micron)
CPU > TR 1950X [email protected]1.2v & [email protected]1.365v[105Mhz FSB](3.23Beta)
#Main: 1700X [email protected]1.365v1.73Ghz IMC CL14 (64.2ns)」+ Dark Rock 3 | X370 Taichi 1.0.0.3 ABBA by The Stilt | Viper Steel [PVS416G400C9K]| Gigabyte 970 G1 | Samsung 850 Evo

Last edited by Veii; 03-26-2020 at 02:29 PM.
Veii is offline  
post #7383 of 7419 (permalink) Old 03-26-2020, 04:50 PM
Noctua Fanboy
 
polkfan's Avatar
 
Join Date: Apr 2015
Posts: 1,043
Rep: 16 (Unique: 13)
I can't seem to figure out how to do that in SiSoftware i can just recall from memory that intercore latency is more consistent now and with each run too! Used to be more over the place like 27-29ns now its 27ns in the same CCX basically with each run and across CCX's its for sure again down 1-2ns and more consistent also even crappy memory setups should see some nice gains as even my worst setting was still better then some setups in terms of memory latency 77ns or so heck i seen that with default XMP setups with cheaper hynix kits.

I did try too for some reason you can't even read the results from older setups i think i might be able to import the profiles though


Must be dumb or something lol i can't even get the numbers to show up its like Sisoftware doesn't save all the results Might be easier to just use excel or something lol

Best i could do really

Lowest result is default 3800/1900 timings

Perhaps i will just make some nice google sheets or something and retest myself to make it a bit easier i will do Inter-Core Bandwidth and Inter-Core Latency comparisons. From what i noticed for these 2 tests FCLK+Memory speeds made WAY more of a difference then timings.
Attached Thumbnails
Click image for larger version

Name:	Capture.PNG
Views:	48
Size:	8.8 KB
ID:	335212  

Click image for larger version

Name:	3.PNG
Views:	46
Size:	11.3 KB
ID:	335216  

Click image for larger version

Name:	2.PNG
Views:	45
Size:	11.2 KB
ID:	335218  


3700X PBO (IF, 1900mhz)16GB G.SKILL TridentZ Series Samsung B-die 3800mhz 14-17-10-32-42
MSI GTX 1080 Gaming X
EVGA SuperNOVA 750 G3
Fractal Design R6(All fans replaced with 5 Noctua A14 fans)
H150i(Fans replaced with Noctua NF-A12)

Noctua fanboy detected! Fans all run at 500RPM

Last edited by polkfan; 03-26-2020 at 04:58 PM.
polkfan is offline  
Sponsored Links
Advertisement
 
post #7384 of 7419 (permalink) Old 03-26-2020, 05:43 PM
Stock is too casual~
 
Veii's Avatar
 
Join Date: Aug 2018
Location: Austria / Bulgaria
Posts: 719
Rep: 37 (Unique: 27)
Ment graphs like this

Only the detailed Page, and filter to local results only

SiSandra Inter-Core Bandwith 14nm Ryzen global #1 | Twitter
Personal Records
RAM > Viper Steel @ 3466C14 (64.2ns) & Alt | Hynix-MFR 3333CL14 (72.5ns) & Alt

GPU > Gigabyte 970 G1 @ 1595|1900 (Hynix) | RX 580 Red Devil GS @ 1620|2200 (Micron)
CPU > TR 1950X [email protected]1.2v & [email protected]1.365v[105Mhz FSB](3.23Beta)
#Main: 1700X [email protected]1.365v1.73Ghz IMC CL14 (64.2ns)」+ Dark Rock 3 | X370 Taichi 1.0.0.3 ABBA by The Stilt | Viper Steel [PVS416G400C9K]| Gigabyte 970 G1 | Samsung 850 Evo
Veii is offline  
post #7385 of 7419 (permalink) Old 03-27-2020, 03:29 PM
Noctua Fanboy
 
polkfan's Avatar
 
Join Date: Apr 2015
Posts: 1,043
Rep: 16 (Unique: 13)
Here is the SiSoftware comparisons for the run i also got 93.3GB when i locked the frequency to 4225mhz.
Attached Thumbnails
Click image for larger version

Name:	3.PNG
Views:	8
Size:	80.3 KB
ID:	335312  

Click image for larger version

Name:	2.PNG
Views:	7
Size:	149.9 KB
ID:	335314  

Click image for larger version

Name:	Capture.PNG
Views:	8
Size:	185.7 KB
ID:	335316  


3700X PBO (IF, 1900mhz)16GB G.SKILL TridentZ Series Samsung B-die 3800mhz 14-17-10-32-42
MSI GTX 1080 Gaming X
EVGA SuperNOVA 750 G3
Fractal Design R6(All fans replaced with 5 Noctua A14 fans)
H150i(Fans replaced with Noctua NF-A12)

Noctua fanboy detected! Fans all run at 500RPM
polkfan is offline  
post #7386 of 7419 (permalink) Old 03-27-2020, 03:38 PM
Stock is too casual~
 
Veii's Avatar
 
Join Date: Aug 2018
Location: Austria / Bulgaria
Posts: 719
Rep: 37 (Unique: 27)
The last picture with different written out results is what i was looking for
Meanwhile both results look awkward
seems like our Blue result is throttled, while our green result was just bad to begin with

Something between the CAD_BUS and proc combination is not fine
Or voltages somewhere are too high and throttled

SiSandra Inter-Core Bandwith 14nm Ryzen global #1 | Twitter
Personal Records
RAM > Viper Steel @ 3466C14 (64.2ns) & Alt | Hynix-MFR 3333CL14 (72.5ns) & Alt

GPU > Gigabyte 970 G1 @ 1595|1900 (Hynix) | RX 580 Red Devil GS @ 1620|2200 (Micron)
CPU > TR 1950X [email protected]1.2v & [email protected]1.365v[105Mhz FSB](3.23Beta)
#Main: 1700X [email protected]1.365v1.73Ghz IMC CL14 (64.2ns)」+ Dark Rock 3 | X370 Taichi 1.0.0.3 ABBA by The Stilt | Viper Steel [PVS416G400C9K]| Gigabyte 970 G1 | Samsung 850 Evo
Veii is offline  
post #7387 of 7419 (permalink) Old 03-27-2020, 05:55 PM
Noctua Fanboy
 
polkfan's Avatar
 
Join Date: Apr 2015
Posts: 1,043
Rep: 16 (Unique: 13)
The tune settings are with the fast profile on 3800mhz in 1.7 dram calc.

Issue with SiSoftware is even the slightest change in the background can have a big impact in the results i noticed. Also locking the frequency does keep it a bit more consistent.

3700X PBO (IF, 1900mhz)16GB G.SKILL TridentZ Series Samsung B-die 3800mhz 14-17-10-32-42
MSI GTX 1080 Gaming X
EVGA SuperNOVA 750 G3
Fractal Design R6(All fans replaced with 5 Noctua A14 fans)
H150i(Fans replaced with Noctua NF-A12)

Noctua fanboy detected! Fans all run at 500RPM
polkfan is offline  
post #7388 of 7419 (permalink) Old 03-28-2020, 06:53 AM
New to Overclock.net
 
brenopapito's Avatar
 
Join Date: May 2016
Location: São Paulo
Posts: 98
Rep: 2 (Unique: 2)
@Veii

Can you make a mini guide with the latest findings on memory overclocking? I have a X470 Taichi with 3.92 bios and I can't reach stability @ 3733.

| Lian Li O11 | Ryzen 3700X | X470 Taichi | GTX 1080 Ti Poseidon | Trident Z RGB 16GB 3600MHz | HX1000i | 970 EVO Plus 500GB | Trion 150 480GB | Water Cooler Custom | Asus PG278QR | Ducky One 2 TKL | Roccat Kone Pure Owl-Eye | Steelseries QcK |
brenopapito is offline  
post #7389 of 7419 (permalink) Old 03-28-2020, 10:43 AM
New to Overclock.net
 
Join Date: Dec 2018
Posts: 369
Rep: 6 (Unique: 6)
what is up with the Special JZ UEFI?
any difference between 6.20, 6.20A, 6.20B?

Ryzen gen.3
(6 items)
CPU
Ryzen 5 3600
Motherboard
ASRock X370 Taichi
GPU
Palit GTX 1080 GameRock
RAM
Corsair Vengeance LPX
Hard Drive
Intel 760p
Case
Fractal Design Define R5
▲ hide details ▲
garych is offline  
post #7390 of 7419 (permalink) Old 03-28-2020, 11:09 AM
New to Overclock.net
 
iNeri's Avatar
 
Join Date: May 2014
Location: México
Posts: 670
Rep: 7 (Unique: 7)
Quote: Originally Posted by garych View Post
what is up with the Special JZ UEFI?
any difference between 6.20, 6.20A, 6.20B?
https://www.overclock.net/forum/28381642-post7351.html

Ryzen [email protected] bclk+ Kraken X61 push/pull|Asrock X370 Taichi|EVGA SC2 1080ti [email protected] & [email protected] Mhz 0.98 mV + Kraken X61|GSkill FlareX 8GBx2 3740MHz CL 16-15-14-14-28-42-282-1t&1.39 mV|EVGA SuperNova 850W|Thermaltake View 31 RGB|NVME CorsairMP500 240 GB+Seagate 3TB 7200rpm 200 MBps|HP Omen 27" [email protected]
iNeri is offline  
Reply

Quick Reply
Message:
Options

Register Now

In order to be able to post messages on the Overclock.net - An Overclocking Community forums, you must first register.
Please enter your desired user name, your email address and other required details in the form below.
User Name:
If you do not want to register, fill this field only and the name will be used as user name for your post.
Password
Please enter a password for your user account. Note that passwords are case-sensitive.
Password:
Confirm Password:
Email Address
Please enter a valid email address for yourself.
Email Address:

Log-in



Currently Active Users Viewing This Thread: 1 (0 members and 1 guests)
 
Thread Tools
Show Printable Version Show Printable Version
Email this Page Email this Page


Forum Jump: 

Posting Rules  
You may post new threads
You may post replies
You may not post attachments
You may not edit your posts

BB code is On
Smilies are On
[IMG] code is On
HTML code is Off
Trackbacks are Off
Pingbacks are Off
Refbacks are Off