Jump to content

"Normal" vs T-series Intel CPUs

Zmimgo

greetings all tech nerds who sit at their computer all day looking at useless graphs on how a large blue boi and a smaller red boi fight with their compressed sand and silicon that can count what 9+10 is

 

So basically, a project I am now starting to work on.

 

Want to make a cheap yet performant NAS/Home server with virtual machines, and am currently looking into what CPU to get. I want to have the system be as power efficient as possible while still being performant enough, and having it in one single system, both the VMs and the NAS (because space). Considoring that the NAS won't be be active at all times, I have been looking for idle power consumption but information about that is minimal on the internet. 

 

What I am wondering is the difference between idle state power draw between something like the i5-11400 and the i5-11400T, and also the Ryzen 5 Pro 4650G if anyone has any data on that. Or if I shoukd just get some old Xeon E5 instead of those CPUs with "only 6 cores?!" The price for electricity here is around 0,38 swedish kronor for every kilowatt, and I'll be buying the parts on the chinese second hand market.

 

thanks

Link to comment
Share on other sites

Link to post
Share on other sites

27 minutes ago, Zmimgo said:

looking at useless graphs on how a large blue boi and a smaller red boi fight with their compressed sand and silicon that can count what 9+10 is

Those graphs aren't useless. Mostly.

 

28 minutes ago, Zmimgo said:

What I am wondering is the difference between idle state power draw between something like the i5-11400 and the i5-11400T,

The -T skews are designed specifically around being low power, though their biggest disadvantage is that because of that, their single core performance isn't really that great in comparison to their non-T skew. What are you trying to run? If it's just a NAS on gigabit ethernet with a Plex server and maybe a discord bot, then the -T skew is probably fine. 

 

31 minutes ago, Zmimgo said:

also the Ryzen 5 Pro 4650G if anyone has any data on that.

I don't personally have any data on that when it comes to power draw, though the biggest advantage of that chip is that it supports ECC memory. Depending on how you're gonna configure your NAS, that might be something that would be very important (I.E. if you'll be using ZFS as the filesystem). 

 

34 minutes ago, Zmimgo said:

Or if I shoukd just get some old Xeon E5 instead of those CPUs with "only 6 cores?!"

That's what I'd go for if you can find a complete server box for cheap. You can find C602 stuff (Xeon E5 V1 and V2) for ridiculously cheap if you know where to look, and C612 stuff (Xeon E5 V3 and V4) for not much more. For example, in my local area, someone is selling a single socket C602 system for $50 US, and I personally bought a dual socket Xeon E5 2630 V1 system with 64GB of ECC RAM for $125 US complete with 8 hot swap drive bays. Plus then you get cool features like IPMI, redundant everything, hot swap everything, rack mountable potential, etc. Plus, you're paying little enough in power (if my conversion is correct, I don't know much about the swedish market) that making up the difference in cost would take a while. Unless what you do requires single thread muscle (I.E. Minecraft servers) then this is probably your best bet.

 

 

Link to comment
Share on other sites

Link to post
Share on other sites

1 hour ago, Zmimgo said:

What I am wondering is the difference between idle state power draw between something like the i5-11400 and the i5-11400T, and also

Its basically the same, the t models are made for when tdp is limited, like in sff systems.

 

If you want lower idle power, the board and other parts make a much bigger difference.

 

1 hour ago, Zmimgo said:

Or if I shoukd just get some old Xeon E5 instead of those CPUs with "only 6 cores?!" The price for electricity here is around 0,38 swedish kronor for every kilowatt, and I'll be buying the parts on the chinese second hand market.

These systems will use much more power at idle depenting on the exact setup. Id go with the 11400 here, faster for most uses, much less power. If you do go this route id stay away from the new chinese boards, and get something from supermicro or simmilar thats used. There much better quality boards that work correctly.

Link to comment
Share on other sites

Link to post
Share on other sites

1 hour ago, RONOTHAN## said:

Those graphs aren't useless. Mostly.

was a joke

 

1 hour ago, RONOTHAN## said:

the biggest advantage of that chip is that it supports ECC memory.

Thought of that as well, but ECC is also a bit more expensive, and it would limit me to AMD.

 

37 minutes ago, Electronics Wizardy said:

Its basically the same, the t models are made for when tdp is limited, like in sff systems.

 

If you want lower idle power, the board and other parts make a much bigger difference.

Yeah, comfirmed what I thought. Thanks

 

38 minutes ago, Electronics Wizardy said:

These systems will use much more power at idle depenting on the exact setup. Id go with the 11400 here, faster for most uses, much less power.

Same reason why I even considored modern processors in the first place. But because of my use of virtual machines, I even considoring i7/i9 Extreme editions or Threadrippers.

 

Thanks for your opinions, now I just need to find a case that can fit 4 HDDs and isn't shit (and doesn't cost more than 50 dollars)

Link to comment
Share on other sites

Link to post
Share on other sites

I have some experience to share. I had a 3900x + X570-Prime-Pro in my server, that consumed approximately 100W in idle. Switched to a i3-10100 + B560M motherboard, dropped to about 50-60W.

 

General consensus around homelab subreddits/communities seems to be, that Intel has better idle power consumption, whereas AMD has better load power consumption (and better price-to-performance etc).

 

I do have a 4750G in a desktop, but I do not have a power meter at hand, to measure how much that system consumes at idle.

HAL9000: AMD Ryzen 9 3900x | Noctua NH-D15 chromax.black | 32 GB Corsair Vengeance LPX DDR4 3200 MHz | Asus X570 Prime Pro | ASUS TUF 3080 Ti | 1 TB Samsung 970 Evo Plus + 1 TB Crucial MX500 + 6 TB WD RED | Corsair HX1000 | be quiet Pure Base 500DX | LG 34UM95 34" 3440x1440

Hydrogen server: Intel i3-10100 | Cryorig M9i | 64 GB Crucial Ballistix 3200MHz DDR4 | Gigabyte B560M-DS3H | 33 TB of storage | Fractal Design Define R5 | unRAID 6.9.2

Carbon server: Fujitsu PRIMERGY RX100 S7p | Xeon E3-1230 v2 | 16 GB DDR3 ECC | 60 GB Corsair SSD & 250 GB Samsung 850 Pro | Intel i340-T4 | ESXi 6.5.1

Big Mac cluster: 2x Raspberry Pi 2 Model B | 1x Raspberry Pi 3 Model B | 2x Raspberry Pi 3 Model B+

Link to comment
Share on other sites

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

×