Jump to content
Search In
  • More options...
Find results that contain...
Find results in...
TheNaitsyrk

Disabling Hyperthreading raised my scores?

25 minutes ago, TheNaitsyrk said:

I'll try RAM next. Got B-Die Samsung. Will try to bring it to 4000Mhz.

Dont forget about the timings and sub-timings as its more about latency than frequency. That's why I suggested you to disable HT in the first place in your threads of yonder. You should have a proper sit down and actually optimize the system as a whole before blaming one component on a poor score.

 

And most people who know what they're doing do disable HT anyway, just look at the hall of fame, top 4 entries are for a 18 logical cored from a 9/7980xe

Recommended Posts

Posted · Original PosterOP

Hiho,

 

Had issues with my 7980XE. I know for sure it's not faulty but 3D Mark is still refusing to work right.

 

Now I decided to disable hyperthreading and my CPU scores went up massively. Can someone suggest what to do to make hyperthreading work right and why it's not working correctly now?

 

The one with low FPS is at 4.7Ghz with HP enabled

The one with sub 200 FPS is 4.7Ghz with HP disabled 

received_477450583134093.jpeg

received_648114369334406.jpeg


Main PC:

CPU: Octodeca Core 7980XE 4.7Ghz OC 1.28V (All threads)

7980XE 4.9Ghz OC 1.30V (HT Disabled)

https://www.3dmark.com/3dm/44422924

Motherboard: EVGA X299 Dark. 

RAM: Corsair Vengeance 16GB x 4 DIMMS, C14 3200Mhz, OC to 3800Mhz C16. Samsung B-Die.

GPU: Water cooled RTX 2080 Ti Founders Edition SLI (+170 Core) (+930 memory) (2145 boost clock on both) with EKWB Vector blocks on each.

Case: ThermalTake X9.

Storage: 2x Samsung EVO 970 1TB NVMe (3500MB/S Read, 3300MB/S Write), 2x Samsung EVO 860 500GB SSD, 1x Samsung EVO 860 4TB SSD, 2x 2TB Seagate Barracuda.

PSU: Corsair AX1600i with custom black and red cables.

Display: 3x Ultrawide Acer Predator X34P 34" Inch G-Sync 120Hz displays IPS with Nvidia surround running at 10320 x 1440 120Hz.

Cooling: Custom Loop EKWB CoolStream 480mm XE 3x radiators and 1x EKWB CoolStream 240mm XE radiator with 2x D5 Revo RGB PUMP, 14x Noctua Industrial 3000RPM fans on rads and 2x in the case, 200mm ThermalTake 800RPM fan (front).

VR: Pimax 4K with NOLO VR.

Keyboard: Corsair K70 LUX RGB.

Sound: T6300 Creative 5.1 speakers.

Mouse: Logitech G502 Proteus Spectrum.

Piano: Yamaha P155.

Phone: Razer Phone 2.

Smart watch: Moto 360 2nd Gen.

Camera: Logitech C922.

3D Printer: Creality Ender 3 Pro.

Link to post
Share on other sites

You increase per core IPC shutting off HT. Since Time spy doesnt scale that many threads, your score increased. 

Nifty aint it?


 

Lid-less PGA 2700x / ROG B450-I Gaming / G.Skill 3600mhz B-Die / GTX 980 bmod / Antec 850w CP series

Lid-less 8700K / Maximus X Hero / G.Skill 4266mhz B-Die / VGA any benching rig / Antec 1000w CP series

Ryzen 5 1400 / Asus Prime A320M-K / Corsair LPX 3000mhz / GTX-980 Bios modded / EVGA 750 / HTPC

Ryzen Athlon 220ge with Vega Graphics / Asus Prime B450M-A / Corsair LED 3000mhz / 550w Antec Office PC.

DFI LanParty UT / 4000+ ClawHammer / DDR Corsair XMS Expert / X1800 XT 256MB Bios modded

Asus CrossHair Formula IV / Phenom II x4 B97 / Dominator GT 2000mhz. / EVGA GTX 770

Link to post
Share on other sites
Posted · Original PosterOP
5 minutes ago, ShrimpBrime said:

You increase per core IPC shutting off HT. Since Time spy doesnt scale that many threads, your score increased. 

Nifty aint it?

Incorrect. At 4.7Ghz with HT on I should be getting 15000+ Physics, I'm getting 12200.

Disabling HT increased it by 1.5K for literally no reason.

 

Proof: https://www.3dmark.com/spy/7984223

 

This person's HT is not disabled. Only 4.6Ghz.


Main PC:

CPU: Octodeca Core 7980XE 4.7Ghz OC 1.28V (All threads)

7980XE 4.9Ghz OC 1.30V (HT Disabled)

https://www.3dmark.com/3dm/44422924

Motherboard: EVGA X299 Dark. 

RAM: Corsair Vengeance 16GB x 4 DIMMS, C14 3200Mhz, OC to 3800Mhz C16. Samsung B-Die.

GPU: Water cooled RTX 2080 Ti Founders Edition SLI (+170 Core) (+930 memory) (2145 boost clock on both) with EKWB Vector blocks on each.

Case: ThermalTake X9.

Storage: 2x Samsung EVO 970 1TB NVMe (3500MB/S Read, 3300MB/S Write), 2x Samsung EVO 860 500GB SSD, 1x Samsung EVO 860 4TB SSD, 2x 2TB Seagate Barracuda.

PSU: Corsair AX1600i with custom black and red cables.

Display: 3x Ultrawide Acer Predator X34P 34" Inch G-Sync 120Hz displays IPS with Nvidia surround running at 10320 x 1440 120Hz.

Cooling: Custom Loop EKWB CoolStream 480mm XE 3x radiators and 1x EKWB CoolStream 240mm XE radiator with 2x D5 Revo RGB PUMP, 14x Noctua Industrial 3000RPM fans on rads and 2x in the case, 200mm ThermalTake 800RPM fan (front).

VR: Pimax 4K with NOLO VR.

Keyboard: Corsair K70 LUX RGB.

Sound: T6300 Creative 5.1 speakers.

Mouse: Logitech G502 Proteus Spectrum.

Piano: Yamaha P155.

Phone: Razer Phone 2.

Smart watch: Moto 360 2nd Gen.

Camera: Logitech C922.

3D Printer: Creality Ender 3 Pro.

Link to post
Share on other sites

This is clearly something other than just CPU frequency, many people below you frequency have higher scores, and plenty people with higher frequencies have lower scores. Have you played with other things like mesh/cache and ram timings yet?

 

https://www.3dmark.com/newsearch#advanced?test=spy P&cpuId=2267&gpuId=1209&gpuCount=0&deviceType=ALL&memoryChannels=0&country=&scoreType=overallScore&hofMode=false&showInvalidResults=false&freeParams=

Link to post
Share on other sites

Looks like the 3d score increased and the physics score decreased? Am I reading it correctly?


 

Lid-less PGA 2700x / ROG B450-I Gaming / G.Skill 3600mhz B-Die / GTX 980 bmod / Antec 850w CP series

Lid-less 8700K / Maximus X Hero / G.Skill 4266mhz B-Die / VGA any benching rig / Antec 1000w CP series

Ryzen 5 1400 / Asus Prime A320M-K / Corsair LPX 3000mhz / GTX-980 Bios modded / EVGA 750 / HTPC

Ryzen Athlon 220ge with Vega Graphics / Asus Prime B450M-A / Corsair LED 3000mhz / 550w Antec Office PC.

DFI LanParty UT / 4000+ ClawHammer / DDR Corsair XMS Expert / X1800 XT 256MB Bios modded

Asus CrossHair Formula IV / Phenom II x4 B97 / Dominator GT 2000mhz. / EVGA GTX 770

Link to post
Share on other sites
Posted · Original PosterOP
Just now, ShrimpBrime said:

Looks like the 3d score increased and the physics score decreased? Am I reading it correctly?

No HT made my score increase. And it allowed cards to boost? It's mostly GPUs giving the score, and with HT on I get poor FPS.

 

I just managed to do 4.8Ghz without HT and scores went higher also.

 

I'm running 3200Mhz memory and 3D Mark support team said that it's normal that disabling HT increases scores, but why doesn't everyone do it? It baffles me.

 

He said that RAM plays a huge role and that TimeSpy doesn't scale well past 8 cores but it makes no sense other people get 15000 in Physics.

3 minutes ago, For Science! said:

This is clearly something other than just CPU frequency, many people below you frequency have higher scores, and plenty people with higher frequencies have lower scores. Have you played with other things like mesh/cache and ram timings yet?

 

https://www.3dmark.com/newsearch#advanced?test=spy P&cpuId=2267&gpuId=1209&gpuCount=0&deviceType=ALL&memoryChannels=0&country=&scoreType=overallScore&hofMode=false&showInvalidResults=false&freeParams=

I'll try RAM next. Got B-Die Samsung. Will try to bring it to 4000Mhz.


Main PC:

CPU: Octodeca Core 7980XE 4.7Ghz OC 1.28V (All threads)

7980XE 4.9Ghz OC 1.30V (HT Disabled)

https://www.3dmark.com/3dm/44422924

Motherboard: EVGA X299 Dark. 

RAM: Corsair Vengeance 16GB x 4 DIMMS, C14 3200Mhz, OC to 3800Mhz C16. Samsung B-Die.

GPU: Water cooled RTX 2080 Ti Founders Edition SLI (+170 Core) (+930 memory) (2145 boost clock on both) with EKWB Vector blocks on each.

Case: ThermalTake X9.

Storage: 2x Samsung EVO 970 1TB NVMe (3500MB/S Read, 3300MB/S Write), 2x Samsung EVO 860 500GB SSD, 1x Samsung EVO 860 4TB SSD, 2x 2TB Seagate Barracuda.

PSU: Corsair AX1600i with custom black and red cables.

Display: 3x Ultrawide Acer Predator X34P 34" Inch G-Sync 120Hz displays IPS with Nvidia surround running at 10320 x 1440 120Hz.

Cooling: Custom Loop EKWB CoolStream 480mm XE 3x radiators and 1x EKWB CoolStream 240mm XE radiator with 2x D5 Revo RGB PUMP, 14x Noctua Industrial 3000RPM fans on rads and 2x in the case, 200mm ThermalTake 800RPM fan (front).

VR: Pimax 4K with NOLO VR.

Keyboard: Corsair K70 LUX RGB.

Sound: T6300 Creative 5.1 speakers.

Mouse: Logitech G502 Proteus Spectrum.

Piano: Yamaha P155.

Phone: Razer Phone 2.

Smart watch: Moto 360 2nd Gen.

Camera: Logitech C922.

3D Printer: Creality Ender 3 Pro.

Link to post
Share on other sites
Posted · Best Answer
25 minutes ago, TheNaitsyrk said:

I'll try RAM next. Got B-Die Samsung. Will try to bring it to 4000Mhz.

Dont forget about the timings and sub-timings as its more about latency than frequency. That's why I suggested you to disable HT in the first place in your threads of yonder. You should have a proper sit down and actually optimize the system as a whole before blaming one component on a poor score.

 

And most people who know what they're doing do disable HT anyway, just look at the hall of fame, top 4 entries are for a 18 logical cored from a 9/7980xe

Link to post
Share on other sites
Posted · Original PosterOP
39 minutes ago, For Science! said:

Dont forget about the timings and sub-timings as its more about latency than frequency. That's why I suggested you to disable HT in the first place in your threads of yonder. You should have a proper sit down and actually optimize the system as a whole before blaming one component on a poor score.

 

And most people who know what they're doing do disable HT anyway, just look at the hall of fame, top 4 entries are for a 18 logical cored from a 9/7980xe

Okay so, score increased by 1k from going 3200 to 3800Mhz.... I'll just try with HT enabled now to see the difference.

https://www.3dmark.com/spy/10777269


Main PC:

CPU: Octodeca Core 7980XE 4.7Ghz OC 1.28V (All threads)

7980XE 4.9Ghz OC 1.30V (HT Disabled)

https://www.3dmark.com/3dm/44422924

Motherboard: EVGA X299 Dark. 

RAM: Corsair Vengeance 16GB x 4 DIMMS, C14 3200Mhz, OC to 3800Mhz C16. Samsung B-Die.

GPU: Water cooled RTX 2080 Ti Founders Edition SLI (+170 Core) (+930 memory) (2145 boost clock on both) with EKWB Vector blocks on each.

Case: ThermalTake X9.

Storage: 2x Samsung EVO 970 1TB NVMe (3500MB/S Read, 3300MB/S Write), 2x Samsung EVO 860 500GB SSD, 1x Samsung EVO 860 4TB SSD, 2x 2TB Seagate Barracuda.

PSU: Corsair AX1600i with custom black and red cables.

Display: 3x Ultrawide Acer Predator X34P 34" Inch G-Sync 120Hz displays IPS with Nvidia surround running at 10320 x 1440 120Hz.

Cooling: Custom Loop EKWB CoolStream 480mm XE 3x radiators and 1x EKWB CoolStream 240mm XE radiator with 2x D5 Revo RGB PUMP, 14x Noctua Industrial 3000RPM fans on rads and 2x in the case, 200mm ThermalTake 800RPM fan (front).

VR: Pimax 4K with NOLO VR.

Keyboard: Corsair K70 LUX RGB.

Sound: T6300 Creative 5.1 speakers.

Mouse: Logitech G502 Proteus Spectrum.

Piano: Yamaha P155.

Phone: Razer Phone 2.

Smart watch: Moto 360 2nd Gen.

Camera: Logitech C922.

3D Printer: Creality Ender 3 Pro.

Link to post
Share on other sites
Posted · Original PosterOP

Okay, HT enabled causes same issues. Score slightly higher with 3800 memory.

 

I guess it's sorted then, but I still don't understand why people got 15000 CPU score with HT enabled....


Main PC:

CPU: Octodeca Core 7980XE 4.7Ghz OC 1.28V (All threads)

7980XE 4.9Ghz OC 1.30V (HT Disabled)

https://www.3dmark.com/3dm/44422924

Motherboard: EVGA X299 Dark. 

RAM: Corsair Vengeance 16GB x 4 DIMMS, C14 3200Mhz, OC to 3800Mhz C16. Samsung B-Die.

GPU: Water cooled RTX 2080 Ti Founders Edition SLI (+170 Core) (+930 memory) (2145 boost clock on both) with EKWB Vector blocks on each.

Case: ThermalTake X9.

Storage: 2x Samsung EVO 970 1TB NVMe (3500MB/S Read, 3300MB/S Write), 2x Samsung EVO 860 500GB SSD, 1x Samsung EVO 860 4TB SSD, 2x 2TB Seagate Barracuda.

PSU: Corsair AX1600i with custom black and red cables.

Display: 3x Ultrawide Acer Predator X34P 34" Inch G-Sync 120Hz displays IPS with Nvidia surround running at 10320 x 1440 120Hz.

Cooling: Custom Loop EKWB CoolStream 480mm XE 3x radiators and 1x EKWB CoolStream 240mm XE radiator with 2x D5 Revo RGB PUMP, 14x Noctua Industrial 3000RPM fans on rads and 2x in the case, 200mm ThermalTake 800RPM fan (front).

VR: Pimax 4K with NOLO VR.

Keyboard: Corsair K70 LUX RGB.

Sound: T6300 Creative 5.1 speakers.

Mouse: Logitech G502 Proteus Spectrum.

Piano: Yamaha P155.

Phone: Razer Phone 2.

Smart watch: Moto 360 2nd Gen.

Camera: Logitech C922.

3D Printer: Creality Ender 3 Pro.

Link to post
Share on other sites
Posted · Original PosterOP

Okay, so memory, disabling HT and OC to 4.9Ghz did the trick.

 

https://www.3dmark.com/3dm/44280990

 

Thanks for your help guys.


Main PC:

CPU: Octodeca Core 7980XE 4.7Ghz OC 1.28V (All threads)

7980XE 4.9Ghz OC 1.30V (HT Disabled)

https://www.3dmark.com/3dm/44422924

Motherboard: EVGA X299 Dark. 

RAM: Corsair Vengeance 16GB x 4 DIMMS, C14 3200Mhz, OC to 3800Mhz C16. Samsung B-Die.

GPU: Water cooled RTX 2080 Ti Founders Edition SLI (+170 Core) (+930 memory) (2145 boost clock on both) with EKWB Vector blocks on each.

Case: ThermalTake X9.

Storage: 2x Samsung EVO 970 1TB NVMe (3500MB/S Read, 3300MB/S Write), 2x Samsung EVO 860 500GB SSD, 1x Samsung EVO 860 4TB SSD, 2x 2TB Seagate Barracuda.

PSU: Corsair AX1600i with custom black and red cables.

Display: 3x Ultrawide Acer Predator X34P 34" Inch G-Sync 120Hz displays IPS with Nvidia surround running at 10320 x 1440 120Hz.

Cooling: Custom Loop EKWB CoolStream 480mm XE 3x radiators and 1x EKWB CoolStream 240mm XE radiator with 2x D5 Revo RGB PUMP, 14x Noctua Industrial 3000RPM fans on rads and 2x in the case, 200mm ThermalTake 800RPM fan (front).

VR: Pimax 4K with NOLO VR.

Keyboard: Corsair K70 LUX RGB.

Sound: T6300 Creative 5.1 speakers.

Mouse: Logitech G502 Proteus Spectrum.

Piano: Yamaha P155.

Phone: Razer Phone 2.

Smart watch: Moto 360 2nd Gen.

Camera: Logitech C922.

3D Printer: Creality Ender 3 Pro.

Link to post
Share on other sites

Create an account or sign in to comment

You need to be a member in order to leave a comment

Create an account

Sign up for a new account in our community. It's easy!

Register a new account

Sign in

Already have an account? Sign in here.

Sign In Now


×