Jump to content

I’m after some advice. Could somebody please suggest the fastest currently available single thread speed cpu.

The cpu needs to have at least 8 cores to run 8x single thread EDA CAD simulations.

overclocking is not an option.

 

i’m thinking 9900ks (best IPC and max clock speed combination?) with the fully sustainable 5Ghz on all 8 cores, paired with the fastest memory that cpu supports. (Configured in dual channel).

 

Am I missing something here? Is anything from AMD in the same ballpark?

Link to comment
https://linustechtips.com/topic/1153917-fastest-single-thread-speed/
Share on other sites

Link to post
Share on other sites

5 minutes ago, Lightning_fingers said:

I’m after some advice. Could somebody please suggest the fastest currently available single thread speed cpu.

The cpu needs to have at least 8 cores to run 8x single thread EDA CAD simulations.

overclocking is not an option.

 

i’m thinking 9900ks (best IPC and max clock speed combination?) with the fully sustainable 5Ghz on all 8 cores, paired with the fastest memory that cpu supports. (Configured in dual channel).

 

Am I missing something here? Is anything from AMD in the same ballpark?

That is the fastest single core chip, yes. For what your specifically asking, that is the correct chip.

 

But why is that the requirement, and why is overclocking not an option? Also, if you are doing CAD sims on each, does RAM bandwidth matter? If so, I’d be very curious if a chip with quad channel RAM would actually net better results. We have no info an ram bandwidth needs in this assumption. 

Rig: i7 13700k +Contact Frame - - Asus Z790-P Wifi - - RTX 4080 - - 4x16GB 6000MHz - - Samsung 990 Pro 2TB NVMe Boot + Main Programs - - Crucial P3 2TB NVMe for photo work - - Corsair RM850x - - Sound BlasterX EA-5 - - Corsair XC8 JTC Edition - - Corsair GPU Full Cover GPU Block - - PTM 7950 - - XT45 X-Flow 420 + UT60 280 rads externally mounted - - EK XRES RGB PWM - - Fractal Define S2 - - DellAlienware AW3423DWF 34" -- Logitech Pro X Superlight - - Logitech G710+ - - LTT Northern Lights Deskpad

 

Headphones/amp/dac: Schiit Bifrost Multibit - -  Schiit Lyr 3 - - Fostex TR-X00 - - Sennheiser HD 6xx

 

Homelab/Media Server: Proxmox VE host - - 512 NVMe Samsung 980 RAID Z1 for VM's/Proxmox boot - - Xeon e5 2660 V4- - Supermicro X10SRF-i - - 128 GB ECC 2133 - - 10x8TB WD Red RAID Z2 - - 2x 800 GB SAS SSD’s (1 SLOG, 1 L2Arc) - - 45 HomeLab HL15 15 Drive 4U - - Corsair RM650i - - LSI 9305-16i HBA - - TreuNAS + many other VM’s

 

Unifi UDM Pro in front of full unifi network infrastructure

 

iPhone 17 Pro - - MacBook Air M3

Link to post
Share on other sites

The 9900KS actually has worse IPC than the 9900K because of mitigations to Intel vulnerabilities. If you want the fastest chip, it's the 9900K with some serious cooling. You can buy pre binned 9900K CPUs as well.

I WILL find your ITX build thread, and I WILL recommend the SIlverstone Sugo SG13B

 

Primary PC:

i7 8086k - EVGA Z370 Classified K - G.Skill Trident Z RGB - WD SN750 - Jedi Order Titan Xp - Hyper 212 Black (with RGB Riing flair) - EVGA G3 650W - dual booting Windows 11 and Fedora Linux - Black and green theme, Razer brainwashed me.

Draws 400 watts under max load, for reference.

 

PSU tier list

How many watts do I need?

PSU misconceptions, protections explainedgroup reg is bad

Link to post
Share on other sites

4 minutes ago, LIGISTX said:

That is the fastest single core chip, yes. For what your specifically asking, that is the correct chip.

 

But why is that the requirement, and why is overclocking not an option? Also, if you are doing CAD sims on each, does RAM bandwidth matter? If so, I’d be very curious if a chip with quad channel RAM would actually net better results. We have no info an ram bandwidth needs in this assumption. 

The requirement is to run 8 concurrent single threaded jobs as quickly as possible.

The EDA CAD Linux based software is not IO bound, but you are right, it’s a tough call if the memory bandwidth compared to a quad channel configuration would have an impact on simulation time.

The assumption is that the 8 concurrent jobs are CPU thread bound.

 

ultimately a hardware evaluation may be required.

 

so dual channel memory 9900ks from intel        (Or a quad channel device) Vs which quad channel device from AMD?

note, cost is also a consideration, ultra expensive xeons are not on the table.

Link to post
Share on other sites

2 minutes ago, Fasauceome said:

The 9900KS actually has worse IPC than the 9900K because of mitigations to Intel vulnerabilities. If you want the fastest chip, it's the 9900K with some serious cooling. You can buy pre binned 9900K CPUs as well.

I’ve read the 9900k can’t sustain 5Ghz across all cores, only 2 cores concurrently.

Link to post
Share on other sites

1 minute ago, Lightning_fingers said:

The requirement is to run 8 concurrent single threaded jobs as quickly as possible.

The EDA CAD Linux based software is not IO bound, but you are right, it’s a tough call if the memory bandwidth compared to a quad channel configuration would have an impact on simulation time.

The assumption is that the 8 concurrent jobs are CPU thread bound.

 

ultimately a hardware evaluation may be required.

 

so dual channel memory 9900ks from intel        (Or a quad channel device) Vs which quad channel device from AMD?

note, cost is also a consideration, ultra expensive xeons are not on the table.

The only quad channel chip in your price range will be thread ripper. But those won’t have the same clock speed at all. But, the RAM bandwidth and cache may help and overcome the lack of sheer ghz. Hard to know for sure tho, I’m just totally guessing at the softwares needs. 

Rig: i7 13700k +Contact Frame - - Asus Z790-P Wifi - - RTX 4080 - - 4x16GB 6000MHz - - Samsung 990 Pro 2TB NVMe Boot + Main Programs - - Crucial P3 2TB NVMe for photo work - - Corsair RM850x - - Sound BlasterX EA-5 - - Corsair XC8 JTC Edition - - Corsair GPU Full Cover GPU Block - - PTM 7950 - - XT45 X-Flow 420 + UT60 280 rads externally mounted - - EK XRES RGB PWM - - Fractal Define S2 - - DellAlienware AW3423DWF 34" -- Logitech Pro X Superlight - - Logitech G710+ - - LTT Northern Lights Deskpad

 

Headphones/amp/dac: Schiit Bifrost Multibit - -  Schiit Lyr 3 - - Fostex TR-X00 - - Sennheiser HD 6xx

 

Homelab/Media Server: Proxmox VE host - - 512 NVMe Samsung 980 RAID Z1 for VM's/Proxmox boot - - Xeon e5 2660 V4- - Supermicro X10SRF-i - - 128 GB ECC 2133 - - 10x8TB WD Red RAID Z2 - - 2x 800 GB SAS SSD’s (1 SLOG, 1 L2Arc) - - 45 HomeLab HL15 15 Drive 4U - - Corsair RM650i - - LSI 9305-16i HBA - - TreuNAS + many other VM’s

 

Unifi UDM Pro in front of full unifi network infrastructure

 

iPhone 17 Pro - - MacBook Air M3

Link to post
Share on other sites

2 minutes ago, Lightning_fingers said:

I’ve read the 9900k can’t sustain 5Ghz across all cores, only 2 cores concurrently.

9900ks can. 

Rig: i7 13700k +Contact Frame - - Asus Z790-P Wifi - - RTX 4080 - - 4x16GB 6000MHz - - Samsung 990 Pro 2TB NVMe Boot + Main Programs - - Crucial P3 2TB NVMe for photo work - - Corsair RM850x - - Sound BlasterX EA-5 - - Corsair XC8 JTC Edition - - Corsair GPU Full Cover GPU Block - - PTM 7950 - - XT45 X-Flow 420 + UT60 280 rads externally mounted - - EK XRES RGB PWM - - Fractal Define S2 - - DellAlienware AW3423DWF 34" -- Logitech Pro X Superlight - - Logitech G710+ - - LTT Northern Lights Deskpad

 

Headphones/amp/dac: Schiit Bifrost Multibit - -  Schiit Lyr 3 - - Fostex TR-X00 - - Sennheiser HD 6xx

 

Homelab/Media Server: Proxmox VE host - - 512 NVMe Samsung 980 RAID Z1 for VM's/Proxmox boot - - Xeon e5 2660 V4- - Supermicro X10SRF-i - - 128 GB ECC 2133 - - 10x8TB WD Red RAID Z2 - - 2x 800 GB SAS SSD’s (1 SLOG, 1 L2Arc) - - 45 HomeLab HL15 15 Drive 4U - - Corsair RM650i - - LSI 9305-16i HBA - - TreuNAS + many other VM’s

 

Unifi UDM Pro in front of full unifi network infrastructure

 

iPhone 17 Pro - - MacBook Air M3

Link to post
Share on other sites

Also, why is overclocking not an option? To be fair, won’t get much more out of a 9900ks anyways, 5.2ghz best case realistically. And AMD OC’s pretty crappy across the board.

If you are just using 8 threads, you could turn hyper threading off which can help with overclocking potential.

Rig: i7 13700k +Contact Frame - - Asus Z790-P Wifi - - RTX 4080 - - 4x16GB 6000MHz - - Samsung 990 Pro 2TB NVMe Boot + Main Programs - - Crucial P3 2TB NVMe for photo work - - Corsair RM850x - - Sound BlasterX EA-5 - - Corsair XC8 JTC Edition - - Corsair GPU Full Cover GPU Block - - PTM 7950 - - XT45 X-Flow 420 + UT60 280 rads externally mounted - - EK XRES RGB PWM - - Fractal Define S2 - - DellAlienware AW3423DWF 34" -- Logitech Pro X Superlight - - Logitech G710+ - - LTT Northern Lights Deskpad

 

Headphones/amp/dac: Schiit Bifrost Multibit - -  Schiit Lyr 3 - - Fostex TR-X00 - - Sennheiser HD 6xx

 

Homelab/Media Server: Proxmox VE host - - 512 NVMe Samsung 980 RAID Z1 for VM's/Proxmox boot - - Xeon e5 2660 V4- - Supermicro X10SRF-i - - 128 GB ECC 2133 - - 10x8TB WD Red RAID Z2 - - 2x 800 GB SAS SSD’s (1 SLOG, 1 L2Arc) - - 45 HomeLab HL15 15 Drive 4U - - Corsair RM650i - - LSI 9305-16i HBA - - TreuNAS + many other VM’s

 

Unifi UDM Pro in front of full unifi network infrastructure

 

iPhone 17 Pro - - MacBook Air M3

Link to post
Share on other sites

Just now, LIGISTX said:

The only quad channel chip in your price range will be thread ripper. But those won’t have the same clock speed at all. But, the RAM bandwidth and cache may help and overcome the lack of sheer ghz. Hard to know for sure tho, I’m just totally guessing at the softwares needs. 

The software does not require 64 bit executables (it’s available in both 32 and 64 bit executable variants). And the memory footprint of each single thread simulation is <4gig therefore the 32 bit executables can be used.

on current hardware, no simulation speed improvement is seen between 32 and 64 bit executables.

 

The dual channel vs quad channel memory configurations, married with differences in IPC and max click speed between Intel and AMD make the decision     tricky.

hence my asking.

 

 

Link to post
Share on other sites

2 minutes ago, Lightning_fingers said:

The software does not require 64 bit executables (it’s available in both 32 and 64 bit executable variants). And the memory footprint of each single thread simulation is <4gig therefore the 32 bit executables can be used.

on current hardware, no simulation speed improvement is seen between 32 and 64 bit executables.

 

The dual channel vs quad channel memory configurations, married with differences in IPC and max click speed between Intel and AMD make the decision     tricky.

hence my asking.

 

 

Yea... hard to say. AMD has faster IPC, but intel has more Ghz. This is why gaming is typically better on intel, but just slightly. The 9900ks is the number 1 gaming chip because of this, but I doubt this is at all apples to apples to your application. Plenty of single threaded things do see gains on AMD because of IPC and such, but that also is not always the case at all. It’s a very hard question your posing, and I’m worried there is only one way to definitively find out... 

Rig: i7 13700k +Contact Frame - - Asus Z790-P Wifi - - RTX 4080 - - 4x16GB 6000MHz - - Samsung 990 Pro 2TB NVMe Boot + Main Programs - - Crucial P3 2TB NVMe for photo work - - Corsair RM850x - - Sound BlasterX EA-5 - - Corsair XC8 JTC Edition - - Corsair GPU Full Cover GPU Block - - PTM 7950 - - XT45 X-Flow 420 + UT60 280 rads externally mounted - - EK XRES RGB PWM - - Fractal Define S2 - - DellAlienware AW3423DWF 34" -- Logitech Pro X Superlight - - Logitech G710+ - - LTT Northern Lights Deskpad

 

Headphones/amp/dac: Schiit Bifrost Multibit - -  Schiit Lyr 3 - - Fostex TR-X00 - - Sennheiser HD 6xx

 

Homelab/Media Server: Proxmox VE host - - 512 NVMe Samsung 980 RAID Z1 for VM's/Proxmox boot - - Xeon e5 2660 V4- - Supermicro X10SRF-i - - 128 GB ECC 2133 - - 10x8TB WD Red RAID Z2 - - 2x 800 GB SAS SSD’s (1 SLOG, 1 L2Arc) - - 45 HomeLab HL15 15 Drive 4U - - Corsair RM650i - - LSI 9305-16i HBA - - TreuNAS + many other VM’s

 

Unifi UDM Pro in front of full unifi network infrastructure

 

iPhone 17 Pro - - MacBook Air M3

Link to post
Share on other sites

10980xe is quad channel as well. But I don’t know what speeds it would result in with 8 cores pegged. Google will know tho. 

Rig: i7 13700k +Contact Frame - - Asus Z790-P Wifi - - RTX 4080 - - 4x16GB 6000MHz - - Samsung 990 Pro 2TB NVMe Boot + Main Programs - - Crucial P3 2TB NVMe for photo work - - Corsair RM850x - - Sound BlasterX EA-5 - - Corsair XC8 JTC Edition - - Corsair GPU Full Cover GPU Block - - PTM 7950 - - XT45 X-Flow 420 + UT60 280 rads externally mounted - - EK XRES RGB PWM - - Fractal Define S2 - - DellAlienware AW3423DWF 34" -- Logitech Pro X Superlight - - Logitech G710+ - - LTT Northern Lights Deskpad

 

Headphones/amp/dac: Schiit Bifrost Multibit - -  Schiit Lyr 3 - - Fostex TR-X00 - - Sennheiser HD 6xx

 

Homelab/Media Server: Proxmox VE host - - 512 NVMe Samsung 980 RAID Z1 for VM's/Proxmox boot - - Xeon e5 2660 V4- - Supermicro X10SRF-i - - 128 GB ECC 2133 - - 10x8TB WD Red RAID Z2 - - 2x 800 GB SAS SSD’s (1 SLOG, 1 L2Arc) - - 45 HomeLab HL15 15 Drive 4U - - Corsair RM650i - - LSI 9305-16i HBA - - TreuNAS + many other VM’s

 

Unifi UDM Pro in front of full unifi network infrastructure

 

iPhone 17 Pro - - MacBook Air M3

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

×