X399M Taichi + Threadripper 1900x Build Log - Page 3 - Overclock.net - An Overclocking Community

Forum Jump: 

X399M Taichi + Threadripper 1900x Build Log

Reply
 
Thread Tools
post #21 of 31 (permalink) Old 09-13-2018, 02:07 AM
New to Overclock.net
 
Kana Chan's Avatar
 
Join Date: Jan 2016
Posts: 685
Rep: 16 (Unique: 14)
How's the ram OC? Did you have time to play around with it?

https://www.overclock.net/forum/10-a...s-club-14.html
Looks like it has potential to scale down quite well.
Kana Chan is offline  
Sponsored Links
Advertisement
 
post #22 of 31 (permalink) Old 09-13-2018, 03:08 AM - Thread Starter
Tank destroyer and a god
 
Offler's Avatar
 
Join Date: Dec 2012
Posts: 2,358
Rep: 86 (Unique: 67)
Quote: Originally Posted by Kana Chan View Post
How's the ram OC? Did you have time to play around with it?

https://www.overclock.net/forum/10-a...s-club-14.html
Looks like it has potential to scale down quite well.
So far I took Gskill Xflare, turned on XMP profile 3200MHz 14-14-14-34 1.35v. That was actualy goal at time when I started to think about new build a year ago.

Most of time I tried to get CPU temperature to 68°C or less, and checking how works "usedplatformclock" and "usedplatformtick" in Win10. I found very interesting that LatencyMon reports 70 microseconds (unless some other process kicks in) and how actually the drivers work in this (for me) completely new environment. Those two settings, in combination with HPET did not made system response worse for me = quite the opposite, yet the 3d score benchmarks might got more precise which usually mean some "virtual performance loss" (only in measuring, but no real performance loss).

Atm, uncore part of CPU is running on higher frequency as is the spec, Core is still within "turbo" range at 4GHz, RAM runs at XMP profile + rest is on AUTO. Since its for the first time I use AMI bios and not AWARD in last 15 years i am bit lost in the terminology for all the settings.

The downside of undervolting (and after BIOS upgrade) was that there was slight, but constant performance loss 2-5 percent when CPU was running "just above" Prime 95 Vcore. So I set it to 1.25v...

On all levels (performance, thermals, stability, network, audio quality, OS and driver behavior) the system so far exceeded my expectations, so i am motivated to do only mild overclock - up to 4180MHz on Cores, and 3200MHz on RAM is already the goal.

Thats why i started to think to get R9 Nano, and do a crossfire with R9 FuryX, or to get rid of Watercooling all along and replace FuryX, in order to make future maintenance of the system even less complicated (regardless how incredible the cooling solution as a whole works).

RAM overclocks to 3400 or 3600 may follow, but so far I dont see much reason for that (other than synthetic benchmarks).

--- Building in progress * AMD Threadripper 1900x * Asrock X399M Taichi * Radeon VII * Gskill Xflare / Samsung B-die 3200 14-14-14-32 * Samsung 970 PRO 512gb * Fractal Design Node 804 * Seasonic Prime Ultra 850 Titanium *
--- Desktop * AMD Phenom II x6 1090t @ 3,8GHz * ASUS M5A99FX PRO R2.0 * Gigabyte R9 FuryX * A-Data XPG 2.0 / Elpida Hyper MNH 1866 @ 1600 6-6-6-18 * LSI 9211-8i / Raid 0 / 5x Corsair Force 3 GS * Creative Xfi Fatal1ty * Intel I210-T1 * Steelseries 6Gv2 * Logitech MX518 * Samsung U28E590D *

Last edited by Offler; 09-14-2018 at 02:25 AM.
Offler is offline  
post #23 of 31 (permalink) Old 09-15-2018, 05:23 AM - Thread Starter
Tank destroyer and a god
 
Offler's Avatar
 
Join Date: Dec 2012
Posts: 2,358
Rep: 86 (Unique: 67)
4000Mhz
1,25v
1 hour of Small FFTs in Prime95

Stable, no errors. CPU temps are slightly above the spec, package power consumption is slightly under the spec (partially due slight undervoltage). I can even lower Vcore to 1,225v.
68°C on CPU was achieved after 15 minutes, 70 took another 30 minutes and 71 another 15.

The test ran in ambient temperature of 26°C. I might try different thermal paste (so far used Noctua) and use Arctic Silver 5. Still best paste in my experience, but needs 1-2 weeks curing time.

Motherboard and SYSIN delta from 34 to 37°C indicates good airflow inside the case. CPU on idle sits on 29°C. What concerns me is that CPU vs Motherboard delta goes from -5 to +34. Air inside the case gets warmer only by 3 degrees, CPU gets warmer by 42... It looks like there is a problem with something which is insulating the CPU, but once the stress test is stopped, temps go down 4 degrees per HWMonitor refresh cycle.
Attached Thumbnails
Click image for larger version

Name:	stress.jpg
Views:	62
Size:	1.80 MB
ID:	218382  


--- Building in progress * AMD Threadripper 1900x * Asrock X399M Taichi * Radeon VII * Gskill Xflare / Samsung B-die 3200 14-14-14-32 * Samsung 970 PRO 512gb * Fractal Design Node 804 * Seasonic Prime Ultra 850 Titanium *
--- Desktop * AMD Phenom II x6 1090t @ 3,8GHz * ASUS M5A99FX PRO R2.0 * Gigabyte R9 FuryX * A-Data XPG 2.0 / Elpida Hyper MNH 1866 @ 1600 6-6-6-18 * LSI 9211-8i / Raid 0 / 5x Corsair Force 3 GS * Creative Xfi Fatal1ty * Intel I210-T1 * Steelseries 6Gv2 * Logitech MX518 * Samsung U28E590D *
Offler is offline  
Sponsored Links
Advertisement
 
post #24 of 31 (permalink) Old 09-27-2018, 06:01 AM - Thread Starter
Tank destroyer and a god
 
Offler's Avatar
 
Join Date: Dec 2012
Posts: 2,358
Rep: 86 (Unique: 67)
To add some updates.

Grabbed one Seasonic Prime Ultra 850 Titanium:
https://seasonic.com/prime-ultra-titanium
https://www.overclock.net/forum/1808...-review-4.html

There were few reasons for that.
a) Corsair 650w I used originally, was already 10 years old, it had just 1x 8-pin EPS connector. 1x4 Pin on CPU VRM was not connected.
As I wrote earlier, the system was working fine even under load, yet CPU VRM was getting at least 130 watt less as the VRM design allowed.

b) Coils on FuryX were whining.
Corsair provided a Y-cable to deliver power to the PCI-E cards. Seasonic provided separated 6+2 cables, however the connection on side of PSU has just 6 pins, and on side of PCI-E, 2 pins are attached to the 6pin connector. At the price tag, and fact they used goldplated connectors, i expected 8pin input and 8 wires to the GPU.

Coils are now not whining under normal gaming (when i usually use Vsync), however the whine does appear on 3dMark and MSI Combustor stress test.

c) Effectivity
1090t was rated at 125w, while 1900x is rated at 180W, and I was looking for a way how to lower the "Total cost of ownership" - which in this case means to lower total power consumption.

My older PC has Seasonic M12ii 850w which is a Bronze rated PSU, with efficiency between 80 and 85%.
http://www2.seasonic.com/product/m12ii-850-evo/

The idea is to spare additional 40-60W solely on higher PSU efficiency while gaming, so new system will have in total about same power consumption. The effectivity of new PSU should be always above 90% and in most cases above 94% (EU, with 230v). Some improvement should in in standby mode.

PSU is also mounted in way it takes fresh air from the bottom, outside of the case, so lower temps for PSU should help the effectivity as well.

So far the temps seem to go down (CPU and VRM), voltages are more stable. However my Wattmetter died, and i have to get new one to compare with old system.

I also looked into details regarding HPET, Windows10, overall driver behavior and measurements by LatencyMON. The underlining philosophy is to have smooth system performance. (Some screenshots will follow).

--- Building in progress * AMD Threadripper 1900x * Asrock X399M Taichi * Radeon VII * Gskill Xflare / Samsung B-die 3200 14-14-14-32 * Samsung 970 PRO 512gb * Fractal Design Node 804 * Seasonic Prime Ultra 850 Titanium *
--- Desktop * AMD Phenom II x6 1090t @ 3,8GHz * ASUS M5A99FX PRO R2.0 * Gigabyte R9 FuryX * A-Data XPG 2.0 / Elpida Hyper MNH 1866 @ 1600 6-6-6-18 * LSI 9211-8i / Raid 0 / 5x Corsair Force 3 GS * Creative Xfi Fatal1ty * Intel I210-T1 * Steelseries 6Gv2 * Logitech MX518 * Samsung U28E590D *

Last edited by Offler; 09-27-2018 at 06:37 AM.
Offler is offline  
post #25 of 31 (permalink) Old 09-27-2018, 12:29 PM - Thread Starter
Tank destroyer and a god
 
Offler's Avatar
 
Join Date: Dec 2012
Posts: 2,358
Rep: 86 (Unique: 67)
So the cooling system is finished, summer in Europe is over as well, so i conducted set of short 15 test.

Stress 3: 15 minutes of Prime95 Small FFTs.
If you check GPU temperature at this point its at 26°C so ... almost no heat gets from CPU to GPU. Temperature at Samsung 970 is at 32°C.

Then Idle1 shows temperatures after few minutes of cooldown.
GPU is at 26°C, Samsung 970 at 30°C. So it indicates that heat from CPU travelled a bit to the drive.

Stress 4:
MSI Kombustor is heating up the GPU. Compared to IDLE on VRM, Mainboard and CPU increased some degrees, however CPU had to perform some tasks for the Kombustor, so the values are affected primarily by this. GPU is at 61°C. Samsung 970 is at 31°C so some heat is getting into the system from it.
Attached Thumbnails
Click image for larger version

Name:	stress3.jpg
Views:	18
Size:	790.7 KB
ID:	220442  

Click image for larger version

Name:	idle1.jpg
Views:	21
Size:	264.0 KB
ID:	220444  

Click image for larger version

Name:	stress4.jpg
Views:	20
Size:	858.9 KB
ID:	220446  


--- Building in progress * AMD Threadripper 1900x * Asrock X399M Taichi * Radeon VII * Gskill Xflare / Samsung B-die 3200 14-14-14-32 * Samsung 970 PRO 512gb * Fractal Design Node 804 * Seasonic Prime Ultra 850 Titanium *
--- Desktop * AMD Phenom II x6 1090t @ 3,8GHz * ASUS M5A99FX PRO R2.0 * Gigabyte R9 FuryX * A-Data XPG 2.0 / Elpida Hyper MNH 1866 @ 1600 6-6-6-18 * LSI 9211-8i / Raid 0 / 5x Corsair Force 3 GS * Creative Xfi Fatal1ty * Intel I210-T1 * Steelseries 6Gv2 * Logitech MX518 * Samsung U28E590D *
Offler is offline  
post #26 of 31 (permalink) Old 10-04-2018, 06:12 AM - Thread Starter
Tank destroyer and a god
 
Offler's Avatar
 
Join Date: Dec 2012
Posts: 2,358
Rep: 86 (Unique: 67)
Ok, lets have a look on how drivers perform on Ryzen platform as a whole. The following report was done while system was idle, HPET disabled.

Code:
_________________________________________________________________________________________________________
CONCLUSION
_________________________________________________________________________________________________________
Your system appears to be suitable for handling real-time audio and other tasks without dropouts. 
LatencyMon has been analyzing your system for  0:01:02  (h:mm:ss) on all processors.


_________________________________________________________________________________________________________
SYSTEM INFORMATION
_________________________________________________________________________________________________________
Computer name:                                        IMMOLATORIII
OS version:                                           Windows 7 , 6.1, build: 7600 (x64)
Hardware:                                             ASRock, X399M Taichi
CPU:                                                  AuthenticAMD AMD Ryzen Threadripper 1900X 8-Core Processor 
Logical processors:                                   16
Processor groups:                                     1
RAM:                                                  32650 MB total


_________________________________________________________________________________________________________
CPU SPEED
_________________________________________________________________________________________________________
Reported CPU speed:                                   3992 MHz

Note: reported execution times may be calculated based on a fixed reported CPU speed. Disable variable speed settings like Intel Speed Step and AMD Cool N Quiet in the BIOS setup for more accurate results.

WARNING: the CPU speed that was measured is only a fraction of the CPU speed reported. Your CPUs may be throttled back due to variable speed settings and thermal issues. It is suggested that you run a utility which reports your actual CPU frequency and temperature. 



_________________________________________________________________________________________________________
MEASURED INTERRUPT TO USER PROCESS LATENCIES
_________________________________________________________________________________________________________
The interrupt to process latency reflects the measured interval that a usermode process needed to respond to a hardware request from the moment the interrupt service routine started execution. This includes the scheduling and execution of a DPC routine, the signaling of an event and the waking up of a usermode thread from an idle wait state in response to that event.

Highest measured interrupt to process latency (µs):   105.413753
Average measured interrupt to process latency (µs):   3.916239

Highest measured interrupt to DPC latency (µs):       102.592460
Average measured interrupt to DPC latency (µs):       1.287283


_________________________________________________________________________________________________________
 REPORTED ISRs
_________________________________________________________________________________________________________
Interrupt service routines are routines installed by the OS and device drivers that execute in response to a hardware interrupt signal.

Highest ISR routine execution time (µs):              0.0
Driver with highest ISR routine execution time:       

Highest reported total ISR routine time (%):          0.0
Driver with highest ISR total time:                   

Total time spent in ISRs (%)                          0.0

ISR count (execution time <250 µs):                   0
ISR count (execution time 250-500 µs):                0
ISR count (execution time 500-999 µs):                0
ISR count (execution time 1000-1999 µs):              0
ISR count (execution time 2000-3999 µs):              0
ISR count (execution time >=4000 µs):                 0


_________________________________________________________________________________________________________
REPORTED DPCs
_________________________________________________________________________________________________________
DPC routines are part of the interrupt servicing dispatch mechanism and disable the possibility for a process to utilize the CPU while it is interrupted until the DPC has finished execution.

Highest DPC routine execution time (µs):              104.288577
Driver with highest DPC routine execution time:       tcpip.sys - TCP/IP Driver, Microsoft Corporation

Highest reported total DPC routine time (%):          0.001230
Driver with highest DPC total execution time:         ntoskrnl.exe - NT Kernel & System, Microsoft Corporation

Total time spent in DPCs (%)                          0.002272

DPC count (execution time <250 µs):                   7794
DPC count (execution time 250-500 µs):                0
DPC count (execution time 500-999 µs):                0
DPC count (execution time 1000-1999 µs):              0
DPC count (execution time 2000-3999 µs):              0
DPC count (execution time >=4000 µs):                 0


_________________________________________________________________________________________________________
 REPORTED HARD PAGEFAULTS
_________________________________________________________________________________________________________
Hard pagefaults are events that get triggered by making use of virtual memory that is not resident in RAM but backed by a memory mapped file on disk. The process of resolving the hard pagefault requires reading in the memory from disk while the process is interrupted and blocked from execution.


Process with highest pagefault count:                 none

Total number of hard pagefaults                       0
Hard pagefault count of hardest hit process:          0
Number of processes hit:                              0


_________________________________________________________________________________________________________
 PER CPU DATA
_________________________________________________________________________________________________________
CPU 0 Interrupt cycle time (s):                       0.207364
CPU 0 ISR highest execution time (µs):                0.0
CPU 0 ISR total execution time (s):                   0.0
CPU 0 ISR count:                                      0
CPU 0 DPC highest execution time (µs):                104.288577
CPU 0 DPC total execution time (s):                   0.021575
CPU 0 DPC count:                                      7584
_________________________________________________________________________________________________________
CPU 1 Interrupt cycle time (s):                       0.092252
CPU 1 ISR highest execution time (µs):                0.0
CPU 1 ISR total execution time (s):                   0.0
CPU 1 ISR count:                                      0
CPU 1 DPC highest execution time (µs):                0.0
CPU 1 DPC total execution time (s):                   0.0
CPU 1 DPC count:                                      0
_________________________________________________________________________________________________________
CPU 2 Interrupt cycle time (s):                       0.055347
CPU 2 ISR highest execution time (µs):                0.0
CPU 2 ISR total execution time (s):                   0.0
CPU 2 ISR count:                                      0
CPU 2 DPC highest execution time (µs):                9.839679
CPU 2 DPC total execution time (s):                   0.000029
CPU 2 DPC count:                                      8
_________________________________________________________________________________________________________
CPU 3 Interrupt cycle time (s):                       0.100044
CPU 3 ISR highest execution time (µs):                0.0
CPU 3 ISR total execution time (s):                   0.0
CPU 3 ISR count:                                      0
CPU 3 DPC highest execution time (µs):                0.0
CPU 3 DPC total execution time (s):                   0.0
CPU 3 DPC count:                                      0
_________________________________________________________________________________________________________
CPU 4 Interrupt cycle time (s):                       0.066707
CPU 4 ISR highest execution time (µs):                0.0
CPU 4 ISR total execution time (s):                   0.0
CPU 4 ISR count:                                      0
CPU 4 DPC highest execution time (µs):                53.166333
CPU 4 DPC total execution time (s):                   0.000763
CPU 4 DPC count:                                      186
_________________________________________________________________________________________________________
CPU 5 Interrupt cycle time (s):                       0.094139
CPU 5 ISR highest execution time (µs):                0.0
CPU 5 ISR total execution time (s):                   0.0
CPU 5 ISR count:                                      0
CPU 5 DPC highest execution time (µs):                0.0
CPU 5 DPC total execution time (s):                   0.0
CPU 5 DPC count:                                      0
_________________________________________________________________________________________________________
CPU 6 Interrupt cycle time (s):                       0.069928
CPU 6 ISR highest execution time (µs):                0.0
CPU 6 ISR total execution time (s):                   0.0
CPU 6 ISR count:                                      0
CPU 6 DPC highest execution time (µs):                22.304609
CPU 6 DPC total execution time (s):                   0.000062
CPU 6 DPC count:                                      9
_________________________________________________________________________________________________________
CPU 7 Interrupt cycle time (s):                       0.089346
CPU 7 ISR highest execution time (µs):                0.0
CPU 7 ISR total execution time (s):                   0.0
CPU 7 ISR count:                                      0
CPU 7 DPC highest execution time (µs):                0.0
CPU 7 DPC total execution time (s):                   0.0
CPU 7 DPC count:                                      0
_________________________________________________________________________________________________________
CPU 8 Interrupt cycle time (s):                       0.060058
CPU 8 ISR highest execution time (µs):                0.0
CPU 8 ISR total execution time (s):                   0.0
CPU 8 ISR count:                                      0
CPU 8 DPC highest execution time (µs):                1.162325
CPU 8 DPC total execution time (s):                   0.000001
CPU 8 DPC count:                                      1
_________________________________________________________________________________________________________
CPU 9 Interrupt cycle time (s):                       0.058537
CPU 9 ISR highest execution time (µs):                0.0
CPU 9 ISR total execution time (s):                   0.0
CPU 9 ISR count:                                      0
CPU 9 DPC highest execution time (µs):                0.0
CPU 9 DPC total execution time (s):                   0.0
CPU 9 DPC count:                                      0
_________________________________________________________________________________________________________
CPU 10 Interrupt cycle time (s):                       0.059045
CPU 10 ISR highest execution time (µs):                0.0
CPU 10 ISR total execution time (s):                   0.0
CPU 10 ISR count:                                      0
CPU 10 DPC highest execution time (µs):                0.0
CPU 10 DPC total execution time (s):                   0.0
CPU 10 DPC count:                                      0
_________________________________________________________________________________________________________
CPU 11 Interrupt cycle time (s):                       0.058645
CPU 11 ISR highest execution time (µs):                0.0
CPU 11 ISR total execution time (s):                   0.0
CPU 11 ISR count:                                      0
CPU 11 DPC highest execution time (µs):                0.0
CPU 11 DPC total execution time (s):                   0.0
CPU 11 DPC count:                                      0
_________________________________________________________________________________________________________
CPU 12 Interrupt cycle time (s):                       0.055972
CPU 12 ISR highest execution time (µs):                0.0
CPU 12 ISR total execution time (s):                   0.0
CPU 12 ISR count:                                      0
CPU 12 DPC highest execution time (µs):                0.0
CPU 12 DPC total execution time (s):                   0.0
CPU 12 DPC count:                                      0
_________________________________________________________________________________________________________
CPU 13 Interrupt cycle time (s):                       0.055809
CPU 13 ISR highest execution time (µs):                0.0
CPU 13 ISR total execution time (s):                   0.0
CPU 13 ISR count:                                      0
CPU 13 DPC highest execution time (µs):                40.711423
CPU 13 DPC total execution time (s):                   0.000043
CPU 13 DPC count:                                      2
_________________________________________________________________________________________________________
CPU 14 Interrupt cycle time (s):                       0.064555
CPU 14 ISR highest execution time (µs):                0.0
CPU 14 ISR total execution time (s):                   0.0
CPU 14 ISR count:                                      0
CPU 14 DPC highest execution time (µs):                36.803607
CPU 14 DPC total execution time (s):                   0.000075
CPU 14 DPC count:                                      4
_________________________________________________________________________________________________________
CPU 15 Interrupt cycle time (s):                       0.062474
CPU 15 ISR highest execution time (µs):                0.0
CPU 15 ISR total execution time (s):                   0.0
CPU 15 ISR count:                                      0
CPU 15 DPC highest execution time (µs):                0.0
CPU 15 DPC total execution time (s):                   0.0
CPU 15 DPC count:                                      0
_________________________________________________________________________________________________________
a) Only 3 drivers actually perform ISR calls.
Thats mouse, keyboard and sound card driver. As mentioned earlier in the topic, the board uses kinda mysterious sound chip Realtek 1220. Its native Windows driver does not have bells and whistles, but also its not making any unnecessary ISR calls. The driver provided by the AsRock provides an agent application, which does ISR calls (tasking the CPU) for no good reason.

When it comes to sound devices, my previous Creative X-fi required a lot of processes, to be hosted on system, and it was causing little higher utilization.

b) Nvidia Vs Radeon
The difference between graphic drivers were well documented. Nvidia tends to make a lot of calls, lowering average system response. Data provided on different threads, and measured by LatencyMon were usually on 250 microseconds (usually reported as nvlddmkm.sys). Current Radeon drivers seem to not perform unnecessary ISR/DPC calls (usually measurable via hal.dll).

In this regard, AMD Radeon GPUs are less intrusive when compared to NVidia.

c) TCP-IP spike
Two drivers on the Windows are usually the slowest, and however there is no way around this. One is NDIS.sys and the other is tcpip.sys. Three DPC calls caused highest measured spike of 0.1ms, while total execution time for all three was 0.15 ms. This can be attributed to fact that network devices still work within range of miliseconds, not microsecond - in that case the result looks really good. The result can be improved a little by fine tunning the Intel 211's driver (disabling power saving features, etc). Currently we speak of the slowest device within the system.

d) Storage - NVMe by Samsung
Apparently no page-fault (reading from disk) happened, but the NVMe driver by Samsung performed few operations. The one which took most time was 0.0018ms, which is incredible. Usage of standard Microsoft driver or SATA would not yield such great results.

e) Mouse and Keyboard.
I connected my keyboard via PS/2 port which is available on the board. Its working on "default" refresh rate, which is 80Hz. No tuning applied yet (can go up to 200Hz). Main reason for that is no Shadowing on keys, and dedicated IRQ.

Its worth noting that USB 3.0 is in case of Ryzens directly on CPU die. The latency is then as good as its possible. Also USB 3.0 even by Asmedia on my older mainboard was superior to USB 2.0 in LatencyMON measurements, so my original objections against USB mice is no longer valid.

Mouse is connected via USB 3.0 port and works up to 1000Hz. Its important to mention "up to" because no mouse driver performs polling if there are no data to transfer (when you dont move the mouse) - real refresh rate is therefore variable. On other hand, the higher refresh rate, the higher CPU utilization for the data processing.

f) CPU utilization
Most of the operation during the LatencyMON benchmark was again lumped on Core 0, Thread 0. Its worth to note that over course of 1 minute benchmark, CPU 0 was busy only for 0.1 milisecond (you got it, TCPIP.sys).

g) Network
As mentioned earlier TCPIP.SYS and NDIS.SYS are the slowest part of the OS at the moment. Ethernet NIC versus Wifi did not caused any major differences in measurements. Ping to router was always >1ms (to get more accurate readings, you need a packet analyzer).

So... Even when I had concerns about smoothness of the system when using Threadripper and Windows 10, all these were proven to be invalid. The system runs almost flawlessly. I will take a look on the network driver, if its behavior can be improved. I this regard i can only highly recommend Ryzen and Threadripper (in combination with Radeon GPU) as an ideal platform for people who are interested less in FPS and more in flawless performance.

Edit:
I disabled unused network cards, disabled power efficient networking (so far let offloading on). TCPIP.SYS and NDIS.SYS went down to 0.065 milisecond in the slowest scenario. Still quite slow for a driver, but definitely worth the tuning.
Attached Thumbnails
Click image for larger version

Name:	latency1.jpg
Views:	35
Size:	235.3 KB
ID:	221734  

Click image for larger version

Name:	latency2.jpg
Views:	38
Size:	346.9 KB
ID:	221736  


--- Building in progress * AMD Threadripper 1900x * Asrock X399M Taichi * Radeon VII * Gskill Xflare / Samsung B-die 3200 14-14-14-32 * Samsung 970 PRO 512gb * Fractal Design Node 804 * Seasonic Prime Ultra 850 Titanium *
--- Desktop * AMD Phenom II x6 1090t @ 3,8GHz * ASUS M5A99FX PRO R2.0 * Gigabyte R9 FuryX * A-Data XPG 2.0 / Elpida Hyper MNH 1866 @ 1600 6-6-6-18 * LSI 9211-8i / Raid 0 / 5x Corsair Force 3 GS * Creative Xfi Fatal1ty * Intel I210-T1 * Steelseries 6Gv2 * Logitech MX518 * Samsung U28E590D *

Last edited by Offler; 10-04-2018 at 09:48 AM.
Offler is offline  
post #27 of 31 (permalink) Old 10-23-2018, 02:01 AM - Thread Starter
Tank destroyer and a god
 
Offler's Avatar
 
Join Date: Dec 2012
Posts: 2,358
Rep: 86 (Unique: 67)
Quote: Originally Posted by Kana Chan View Post
How's the ram OC? Did you have time to play around with it?

https://www.overclock.net/forum/10-a...s-club-14.html
Looks like it has potential to scale down quite well.
Followed the guide here:
https://www.overclock.net/forum/11-a...l#post27679996

So far 3600MHz 14-14-14-32 cr1 @ 1.495v, but not stable in stress test.

Anyway OCCT Linpack - 19 minutes, 31 secs @ 4000MHz on CPU and 3600MHz on RAM meant:
71°C on CPU
50°C on VRM
46-49°C on RAM (Coolest to hottest)
37°C Motherboard/TMPIN.

Power draw "from the wall" 270 watt, indicated CPU power draw 174 watts (slightly lower than TDP). I guess the system crashed when stress stopped for a while - CPU VDroop then goes lower and for a moment it could be below stable value. (@ 3200MHz CPU never went below 1.2v, @ 3600MHz it was at 1.194v. Not a big difference, but good as an indication).

Overall result:
Motherboard/TMPIn indicates that aircooling is working as expected. CPU is not running on stock (3.8GHz), but also its not above maximum turbo (4.2GHz), while voltage is below stock, and TDP is slightly lower as well. Expected temperature was therefore below 68°C.

Ambient temperature is however unknown (if it wasnt 21°C it should be fine).

RAM overclocking to 3333MHz was perceivable as certain processes were visibly faster, so the performance scaling is there = Cinebench went from 1748 @ 3200 to 1798 @ 3600 - which is roughly 3%. Memory latency went from 89ns to 70ns which is quite significant.

--- Building in progress * AMD Threadripper 1900x * Asrock X399M Taichi * Radeon VII * Gskill Xflare / Samsung B-die 3200 14-14-14-32 * Samsung 970 PRO 512gb * Fractal Design Node 804 * Seasonic Prime Ultra 850 Titanium *
--- Desktop * AMD Phenom II x6 1090t @ 3,8GHz * ASUS M5A99FX PRO R2.0 * Gigabyte R9 FuryX * A-Data XPG 2.0 / Elpida Hyper MNH 1866 @ 1600 6-6-6-18 * LSI 9211-8i / Raid 0 / 5x Corsair Force 3 GS * Creative Xfi Fatal1ty * Intel I210-T1 * Steelseries 6Gv2 * Logitech MX518 * Samsung U28E590D *

Last edited by Offler; 10-23-2018 at 05:06 AM.
Offler is offline  
post #28 of 31 (permalink) Old 10-28-2018, 12:17 PM - Thread Starter
Tank destroyer and a god
 
Offler's Avatar
 
Join Date: Dec 2012
Posts: 2,358
Rep: 86 (Unique: 67)
Did first full-fledged test on OCCT Linpack. The test lasted 25 minutes, ambient temp was 23°C.
Attached Thumbnails
Click image for larger version

Name:	linpack.jpg
Views:	50
Size:	447.5 KB
ID:	227686  


--- Building in progress * AMD Threadripper 1900x * Asrock X399M Taichi * Radeon VII * Gskill Xflare / Samsung B-die 3200 14-14-14-32 * Samsung 970 PRO 512gb * Fractal Design Node 804 * Seasonic Prime Ultra 850 Titanium *
--- Desktop * AMD Phenom II x6 1090t @ 3,8GHz * ASUS M5A99FX PRO R2.0 * Gigabyte R9 FuryX * A-Data XPG 2.0 / Elpida Hyper MNH 1866 @ 1600 6-6-6-18 * LSI 9211-8i / Raid 0 / 5x Corsair Force 3 GS * Creative Xfi Fatal1ty * Intel I210-T1 * Steelseries 6Gv2 * Logitech MX518 * Samsung U28E590D *
Offler is offline  
post #29 of 31 (permalink) Old 03-22-2019, 03:49 AM - Thread Starter
Tank destroyer and a god
 
Offler's Avatar
 
Join Date: Dec 2012
Posts: 2,358
Rep: 86 (Unique: 67)
Fast forward few months...

I purchased one Radeon VII, about two days after it was released.

Positives:
Nice performance upgrade
Very low power consumption compared to FuryX, expecially on 1080p gaming scenarios. With FuryX system was consuming 300+ watts with Radeon VII its 160 watts (!).

Negatives:
Stock cooler required lapping
Its much more noisy on high-performance scenarios
Since I managed to install FuryX cooler in the right chamber, but i cannot do that with Radeon VII now, the heat from GPU dissipates in the left chamber, heating up RAM and CPU, and incredibly damaging the original airflow setup...

Also It alienates the "little big build" vibe the original setup had...

Of course there is a way how to fix it and thats to mount a watercooling solution, however... it was really the watercooling on FuryX which made me a bit nervous in case of leak.

--- Building in progress * AMD Threadripper 1900x * Asrock X399M Taichi * Radeon VII * Gskill Xflare / Samsung B-die 3200 14-14-14-32 * Samsung 970 PRO 512gb * Fractal Design Node 804 * Seasonic Prime Ultra 850 Titanium *
--- Desktop * AMD Phenom II x6 1090t @ 3,8GHz * ASUS M5A99FX PRO R2.0 * Gigabyte R9 FuryX * A-Data XPG 2.0 / Elpida Hyper MNH 1866 @ 1600 6-6-6-18 * LSI 9211-8i / Raid 0 / 5x Corsair Force 3 GS * Creative Xfi Fatal1ty * Intel I210-T1 * Steelseries 6Gv2 * Logitech MX518 * Samsung U28E590D *
Offler is offline  
post #30 of 31 (permalink) Old 10-22-2019, 03:11 AM - Thread Starter
Tank destroyer and a god
 
Offler's Avatar
 
Join Date: Dec 2012
Posts: 2,358
Rep: 86 (Unique: 67)
Another fast forward.

1. Updates on system input lag latency.
a) I've decided to enable HPET and let Windows 10 manage system timer features as was originally intended. This is however due implementation features against Spectre vulnerabilities.

b) When you open network details, get to the IPv4 protocol, WINS tab and completely disable Netbios.
This feature is driven by netbt.sys and its kept mostly for legacy purposes. If you experience issues, turn it back on. In some cases this driver caused some additional latency issues.

c) In BIOS memory management was CR set to 1t and Gear Down mode disabled.
Overall CPU power draw increased by 5 watts (constantly over few more cycles of testing| in stress test which indicates slightly faster operation. Also I increased volts from 1.212v to 1.256v while CPU now indicates not 150w power consumption but 176w at the same frequency. As indicated by HWinfo, current for CPUs also increased.

It appears that on lower voltages there were few "dead" cycles, as CPUs run out of power during stress test.

d) Display pixel format was set to YCrCb and then back to RGB.
Due some reason overall responsiveness in games visible increased, while image is more smooth.

2. Planned upgrades
Regardess I wanted to get rid of watercooled GPU i probably have to use it anyway with Radeon VII. The card is obviously hindered by hotspots and proper cooling (while modded CPU coolers shows a lot of promise) might improve the performance.

As a result, the airflow for the CPU cooler should get back to normal and system should accummulate less heat from the GPU.

--- Building in progress * AMD Threadripper 1900x * Asrock X399M Taichi * Radeon VII * Gskill Xflare / Samsung B-die 3200 14-14-14-32 * Samsung 970 PRO 512gb * Fractal Design Node 804 * Seasonic Prime Ultra 850 Titanium *
--- Desktop * AMD Phenom II x6 1090t @ 3,8GHz * ASUS M5A99FX PRO R2.0 * Gigabyte R9 FuryX * A-Data XPG 2.0 / Elpida Hyper MNH 1866 @ 1600 6-6-6-18 * LSI 9211-8i / Raid 0 / 5x Corsair Force 3 GS * Creative Xfi Fatal1ty * Intel I210-T1 * Steelseries 6Gv2 * Logitech MX518 * Samsung U28E590D *

Last edited by Offler; 10-22-2019 at 04:11 AM.
Offler 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