It is rated at max 180w TDP just like the 1080..
But it might actually use a tad less..

The difference is probably minimal.

Thanks for information. I just had a look at your host 8574118. You're currently returning work and just over 4 minutes this could be something to do with noise bombs. Out of interest how many tasks are you running at a time and if more than one do you 3 a CPU core for each task? Thank you for information

This host didn't run since a while actually but i will let it run for a bit see the results i get since the last results we're not blc22s ..(this is a dual boot windows/linux machine)
Currently using the lunatics app set to run 2 gpu , max concurrent task is set at 8.. so running 6cpu/2gpu WUs..
I don't know if this is optimal or the correct app to run.. as i mostly don't run windows much anymore.

Thanks. I would run Linux but I find the applications I use are not compatible plus it is not a user-friendly interface for me (more text driven than mouse driven)
on a very quick side note was it a nightmare to set up an dual boot situation & are you using the same drive for all the petitions/swap and file partition?

---

I don't have any 10x0 cards, not yet at least, maybe asap, providing nothing goes wrong here. I might get a 1070Ti blower or maybe a 1080, it depends on which costs less, either an MSI Aero or a PNY Blower, though I like the MSI a bit more.

I'm running 5 wu's, 3 on the cpu and 2 on the Asus 970 Turbo, the 970 is doing each in about 22mins, if I did 1 at a time that would be about 11mins.

And I've left the Turbo at it's stock speed, tonight I increased the hours to get some more heat, the temp is 64F outside.

---

The T1 Trust, PRR T1 Class 4-4-4-4 #5550, 1 of America's First HST's

GPU WUs seems to be running in the 6min54 range when running 2 of them right now..

It is fairly easy to setup actually, i am running Kubuntu (KDE desktop) which is somewhat more similar to windows than the default Gnome desktop in Ubuntu..
I am quite used to dual booting and dealing with multiple partitions, but i would rather recommend having a whole disk for linux instead and just change device to boot from bios instead.
Because when stuff goes wrong, dealing with OS that don't want to boot anymore because of partition changes can be harder to diagnose and repair and will affect both OS.

That said, it is very simple to dual boot from one disk..
- Shrink your windows partition to free up some space from within windows
- Boot from your linux usb install
- Install it on the free space.

I just let the installer do one big partition in the free space . Ubuntu/Kubuntu do not use swap partition anymore, but a swap file just like windows.

I will most likely buy a 2nd smaller SSD just for linux, as it will totally isolate windows and linux from each other.
For example if you want to reinstall windows, it will erase the linux bootloader and you will have to recover it using command lines stuff when using a single disk.

---

Probably the easiest dual boot you can do is from a USB drive. I haven't ran into any problems yet with a 16G USB Stick that been running for 1.5 years.

For reliability, an external USB 3.0 SSD drive would work great, and you don't need to go big at all. If you ever pack 40Gb on a cruncher I'd be surprised.

Then just set your BIOS to boot from USB first, Window Drive second. Then just unplug the USB whenever you want Windows to boot up.

I don't have any 10x0 cards, not yet at least, maybe asap, providing nothing goes wrong here. I might get a 1070Ti blower or maybe a 1080, it depends on which costs less, either an MSI Aero or a PNY Blower, though I like the MSI a bit more.

I'm running 5 wu's, 3 on the cpu and 2 on the Asus 970 Turbo, the 970 is doing each in about 22mins, if I did 1 at a time that would be about 11mins.

And I've left the Turbo at it's stock speed, tonight I increased the hours to get some more heat, the temp is 64F outside.

17° C is a nice temperature where I am from here in New Zealand. In fact I think we were about that temperature today. That's interesting on your 970 are giving time is roughly the same as me I am completing them in about 9 minutes 47 seconds at the moment. 11 tasks completed in 1 hour 47 minutes 40 seconds. I was hoping to get some noise bombs but doesn't look like it is the case, 65 remaining 10 hours 40 minutes. Using a 970 jet stream

---

Thanks Rick and brent for your ideas in regards to Linux.

---

I have EVGA 1080 FTW Gaming cards which do around 12-13 minutes running two tasks at a time. So I guess that's about 6-6.5 minutes a task. Does that sound about right?

---

I have EVGA 1080 FTW Gaming cards which do around 12-13 minutes running two tasks at a time. So I guess that's about 6-6.5 minutes a task. Does that sound about right?

Thank you for your insight those times are quite reasonable especially for 2 tasks. I must say running 1 task is a very nice time Do you notice a major hit an increase running 2 tasks?

---

I have EVGA 1080 FTW Gaming cards which do around 12-13 minutes running two tasks at a time. So I guess that's about 6-6.5 minutes a task. Does that sound about right?

It looks like you are running the Lunatics distro under Windows?

I just looked on the Leaderboard at one of the top Windows + gtx 1080Ti's and some of his gpu tasks are running in the 5-minute area. So without trying to find someone else who is running a gtx 1080 under Windows w/Lunatics I am guessing you are certainly in the right neighborhood even if there MIGHT be a little more you could squeeze out. Or that maybe all you can squeeze out.

It is a very respectable performance.

Tom

---

A proud member of the OFA (Old Farts Association).

I ended up ordering if 1070 TI because there was no 1080 in stock & IT company was unsure whether there was going to be any more. There was a big run on them because people saw the price of the 2080 and 2080 TI and thought 1080 was the next best option I imagine. Not sure when it will be installed.
I will let you know after I return some results. Just a reminder for people reading this I am in New Zealand. Thank you for all the assistance

---

I've been happy with my 1070Ti's. 95% of the performance of a 1080 for $100 less.

---

Seti@Home classic workunits:20,676 CPU time:74,226 hours

A proud member of the OFA (Old Farts Association)

I've been happy with my 1070Ti's. 95% of the performance of a 1080 for $100 less.

Keith I am keen to know are you running a commandline to get more efficiency out of your 1070 TI's? I am running the following

-sbs 1024 -period_iterations_num 30 -spike_fft_thresh 2048 -tune 1 64 1 4 -oclfft_tune_gr 256 -oclfft_tune_lr 16 -oclfft_tune_wg 256 -oclfft_tune_ls 512 -oclfft_tune_bn 32 -oclfft_tune_cw 32

on my 970. You can find the text file located in the Seti folder in the project folder titled "mb_cmdline-8.22_windows_intel__opencl_nvidia_SoG"

---

Well, first take into account I don't run SoG anymore. I am running nothing but the special app in Linux. But I always move my memory clocks back to what they should be running in P0 state too.

I would probably tighten up your command line with a period_iteratations_num=1 to start. Everything else looks pretty much standard and I wouldn't change anything else. But the longer you can stay in the kernel crunching the better.

---

Seti@Home classic workunits:20,676 CPU time:74,226 hours

A proud member of the OFA (Old Farts Association)

also add

-hp

to the commandline

---

Well, first take into account I don't run SoG anymore. I am running nothing but the special app in Linux. But I always move my memory clocks back to what they should be running in P0 state too.

I would probably tighten up your command line with a period_iteratations_num=1 to start. Everything else looks pretty much standard and I wouldn't change anything else. But the longer you can stay in the kernel crunching the better.

Can you please explain by what you mean "but the longer you stay in the kernel crunching the better" what will adding "period_iteratations_num=1" do?

---

also add

-hp

to the commandline

Thanks for the tip. Well running it in high priority cause more heat?

---

Both iterations=1 and hp will make the card work harder so will make more heat obviously. Raistmer did an excellent job of explaining all the parameters and what they do and why in his post at Lunatics.
Some considerations regarding OpenCL MultiBeam app tuning from algorithm view

The very first part of the post explains -period_iterations_num N and what the application is doing in the compute kernel.

[Edit] To answer your question directly, it will try to process all possible periods in single kernel call when set to one iteration. So the fastest because you didn't break up computation into more pieces.

---

Seti@Home classic workunits:20,676 CPU time:74,226 hours

A proud member of the OFA (Old Farts Association)

Both iterations=1 and hp will make the card work harder so will make more heat obviously. Raistmer did an excellent job of explaining all the parameters and what they do and why in his post at Lunatics.
Some considerations regarding OpenCL MultiBeam app tuning from algorithm view

The very first part of the post explains -period_iterations_num N and what the application is doing in the compute kernel.

[Edit] To answer your question directly, it will try to process all possible periods in single kernel call when set to one iteration. So the fastest because you didn't break up computation into more pieces.

Thank you for information I added the command line you recommended I have not noticed any speed increase apart from -9 overflow result which ended in just over 7 minutes. I haven't tried - HP yet because I feel it's going to increase my fan speed to over 40 % it is currently at 38% and a temperature of 71° C

---

-hp stands for high priority mode. IOW, the thread will get highest priority over all other threads the cpu is running. It can have a deleterious effect and cause system lag and apparent lockups because the system can't service other normal running maintenance threads. You are forewarned.

---

Seti@Home classic workunits:20,676 CPU time:74,226 hours

A proud member of the OFA (Old Farts Association)

Well, first take into account I don't run SoG anymore. I am running nothing but the special app in Linux. But I always move my memory clocks back to what they should be running in P0 state too.

I would probably tighten up your command line with a period_iteratations_num=1 to start. Everything else looks pretty much standard and I wouldn't change anything else. But the longer you can stay in the kernel crunching the better.

-hp stands for high priority mode. IOW, the thread will get highest priority over all other threads the cpu is running. It can have a deleterious effect and cause system lag and apparent lockups because the system can't service other normal running maintenance threads. You are forewarned.

I am aware about what the high priority command does. If I am not seeing a speed benefit using the command you mentioned above I am all for processing work faster but if this is not been achieved is there any benefit to using the command line?

---