SMB and CPU

[Fastline]

Probably. If:

• Your only application is file-sharing
• You're only using SMB
• You have relatively few users who'd be accessing at one time
• Highest possible network throughput is your only priority

then go for a 4+ core CPU with the highest clock speed you can find. Hyperthreading will likely be a given in this case.

As a semi-relevant point of comparison, pushing 10 Gbit/sec (9.4 Gbit/sec actual throughput) through iperf3 comes close to maxing out one core of my Xeon Gold 6132. SMB is going to have more i/o to deal with, of course.

Wohoo mate. You clear all my confusions. Thanks again.

Yes, the work is only what you mentioned so going for a single socket CPU. You saved me $$$ with your precious information :)

 

[Fastline]

I would like to point out that this discussion, at least in my perception, is a bit narrow. Perhaps I missed the part where a precise use-case description was done. That would include details on the workload in terms of throughput, IOPS, etc.

Not that the points above are irrelevant. But they seem to focus on one aspect only. At least as important are the underlying disks and their configuration, the network, the clients (hardware and applications), etc. It is also worth noting that those clients actually need to be able to saturate the server (not only in the context of TrueNAS/storage but in general).

Totally get your point. But yes, was focusing on one part to clear the questions i had.

 

[Patrick M. Hausen]

Do you really expect that some community member (!) will show up who

1) owns both configurations
2) has performed extensive benchmarks

?

Nobody knows, for chrissake. Just go for a reasonable price/performance ratio and availability.

Hint: the last extra GHz are frequently way more expensive than a system in the, say, upper third. Equally for RAM and storage capacity ...

 

[Fastline]

Do you really expect that some community member (!) will show up who

1) owns both configurations
2) has performed extensive benchmarks

?

Not at all.

Nobody knows, for chrissake. Just go for a reasonable price/performance ratio and availability.

Gotcha.

Hint: the last extra GHz are frequently way more expensive than a system in the, say, upper third. Equally for RAM and storage capacity ...

Could you please explain what does upper third here means?

 

[Davvo]

Could you please explain what does upper third here means?

He's about the premium of bleeding edge/top of the line vs the top third in the performance scale. In this case, upper third means the group of CPUs that perform better than 2/3 of all comparable chips.

 

[Fastline]

He's about the premium of bleeding edge/top of the line vs the top third in the performance scale. In this case, upper third means the most performing CPUs.

That's what i though. So, would there be a much difference between 3.7GHz and 4.2GHz CPU when using it for the SMB?

 

[Davvo]

So, would there be a much difference between 3.7GHz and 4.2GHz CPU when using it for the SMB?

In a perfect environment, yes. But do you have a perfect environment?
Are you running enterprise-grade NVMe SSDs in mirrors?
Are you using tons of high performance RAM?
Is your network based on high performance, low latency hardware (switches, fiber optic, NICs)?
Is your network configured in order to achieve the maximum performance and lowest latency possible?
Are your clients using high performance, low latency hardware?
Is your data optimized for those speeds?
...

TL;DR: without further considerations, the major difference is likely to be the price; 0.5GHz is a significant improvement, but only you can judge if it's worth it. Your system, your money.

 

[danb35]

would there be a much difference between 3.7GHz and 4.2GHz CPU when using it for the SMB?

I think that was the point of Patrick's comment: nobody really knows. I mean, yes, in theory, all other things being equal and with nothing else limiting performance, there would be a difference of a little over 12%. But all other things are never equal, and there are always other factors that affect performance.

 

[Fastline]

In a perfect environment, yes. But do you have a perfect environment?
Are you running enterprise-grade NVMe SSDs in mirrors?
Are you using tons of high performance RAM?
Is your network based on high performance, low latency hardware (switches, fiber optic, NICs)?
Is your network configured in order to achieve the maximum performance and lowest latency possible?
Are your clients using high performance, low latency hardware?
Is your data optimized for those speeds?
...

TL;DR: without further considerations, the major difference is likely to be the price; 0.5GHz is a significant improvement, but only you can judge if it's worth it. Your system, your money.

Understood.

The price is like 150 USD rise in if i choose 4.2GHz CPU vs 3.7GHz. If that's not going to make a nominal difference or like one cannot even notice, i would go with 3.7GHz CPU.

 

[Fastline]

I think that was the point of Patrick's comment: nobody really knows. I mean, yes, in theory, all other things being equal and with nothing else limiting performance, there would be a difference of a little over 12%. But all other things are never equal, and there are always other factors that affect performance.

Yes, Thanks. All solved now :)

 

[ChrisRJ]

At the end of the day it is about constraints. If the CPU is not the constraint, getting a better one will not change anything to the overall result.

 

[Fastline]

At the end of the day it is about constraints. If the CPU is not the constraint, getting a better one will not change anything to the overall result.

Perfect!

So i got it guys. If everything is all solid, one can notice a significant difference between 3.7GHz and 4.2GHz but if the rests of the components are not perfect, one won't even notice. Cool cool.

 

[Davvo]

The price is like 150 USD rise in if i choose 4.2GHz CPU vs 3.7GHz.

I would totally go with the 3.7GHz CPU if the base clock is the only difference.

 

[Fastline]

I would totally go with the 3.7GHz CPU if the base clock is the only difference.

Unfortunately, its the boost clock. I might have some old CPUs. I'll do some tests to see if i can see in a test environment. That would help me a bit to choose between both.

 

[MrGuvernment]

I run my TrueNAS off a Intel(R) Xeon(R) Silver 4114 CPU @ 2.20GHz , 10 cores, 20 threads. i can max out 10Gbps using iperf, my underlying I/O of spinning rust cant max out close to 10Gbps, so cant test SMB.

 

[Fastline]

I run my TrueNAS off a Intel(R) Xeon(R) Silver 4114 CPU @ 2.20GHz , 10 cores, 20 threads. i can max out 10Gbps using iperf, my underlying I/O of spinning rust cant max out close to 10Gbps, so cant test SMB.

Gotcha