Posts by Keith Myers

1) Message boards : Number crunching : Intel 16 core/32T discusssion (Message 1956403)
Posted 2 hours ago by Profile Keith Myers Special Project $250 donor
Post:
Depends on the board and the case. I have a SSI CEB form factor workstation X99 board and it is the same size as an EATX board. Fit the same standoffs in the case. Normally the only thing you can get in trouble with a EATX form factor motherboard is occluding the pass-through cable grommets in the case from the power supply. If that isn't an issue, most EATX motherboards fit a ATX case as long as there are EATX standoff positions you can use.
2) Message boards : Number crunching : GTX is dead long live RTX !! (Message 1956400)
Posted 2 hours ago by Profile Keith Myers Special Project $250 donor
Post:
I would say if you have the budget for the 2080 or the 2070 and are running something like the 1060 or 1070, there may be a case to upgrade. Absolutely no case for the 2080Ti at its price point.
[Edit] It's a complete sideways move from anyone already running a 1080 Ti to a 2080. No point.
3) Message boards : Number crunching : GTX is dead long live RTX !! (Message 1956397)
Posted 2 hours ago by Profile Keith Myers Special Project $250 donor
Post:
Looks like Phoronix came up with about the same performance numbers compared to Pascal. About 30% improvement in games at 4K for the 2080 Ti compared to the 1080 Ti. I like their frames per second per watt and frames per second per dollar charts. Shows the 2080 Ti slightly better on the frames per second per watt metric but losing soundly on the frames per second per dollar metric to the 1080Ti.

So the new 2080 card will likely help your power bill compared to the the 1080Ti but you will have the same power bill for the 2080Ti as if you were running the Vega64 card. The Turing FE cards run a few degrees cooler than the Pascal FE cards. The AIB board offerings should improve on the thermals of the FE cards as expected by quite a bit.

Phoronix didn't run their compute benchmark suite unfortunately.
4) Message boards : Number crunching : GTX is dead long live RTX !! (Message 1956390)
Posted 3 hours ago by Profile Keith Myers Special Project $250 donor
Post:
That's what I have been waiting for. So, for N-body simulation, likely comparable to the MilkyWay@Home N-body simulation tasks, 50% faster for the 2080 FE over the 1080 FE. For Folding@Home tasks, 30% faster for the 2080 FE over the 1080 FE. For the 2080 Ti over the 1080 Ti, mostly the same 30-50% faster.

On the Geekbench 4 Compute benchmark, the new Turing cards are 200% faster than the Pascal cards. The Geekbench 4 benchmark are normally run with OpenCL primitives so directly comparable to the SoG app. They also can be run with CUDA primitives but I haven't read the article fully yet to see if they ran both the OpenCL and CUDA primitives and combined the scores or if they were just with the default OpenCL.
5) Message boards : Number crunching : My cruncher is not showing up on my account but recieving and sending work .. (Message 1956384)
Posted 3 hours ago by Profile Keith Myers Special Project $250 donor
Post:
If you just put the fifth system online, it takes a day or two before Seti picks up the host.
6) Message boards : Number crunching : GTX is dead long live RTX !! (Message 1956369)
Posted 4 hours ago by Profile Keith Myers Special Project $250 donor
Post:
The only benefit so far in Seti computing using the special app is the the Titan V has 80 SM compute units versus the 28 SM compute units of the 1080Ti. The 2080 Ti has 68 SM compute units. The 2080 has 46 SM compute units. The 2070 will have 36 SM compute units. So the benefit to Seti compute is the scaling of SM units in the newer architectures. Don't know yet whether a Turing SM unit is comparable to a Pascal SM unit with regard to compute. Still looking for the compute performance benchmarks.
7) Message boards : Number crunching : GTX is dead long live RTX !! (Message 1956367)
Posted 4 hours ago by Profile Keith Myers Special Project $250 donor
Post:
OK, I misnomered. I meant the Tensor cores. Nobody in distributed computing is using the Tensor cores. Certainly not Seti.
8) Message boards : Number crunching : GTX is dead long live RTX !! (Message 1956363)
Posted 5 hours ago by Profile Keith Myers Special Project $250 donor
Post:
No, the RT cores aren't and haven't been able to be used for SETI, or any distributed computing project that I know of. The benefit that the Titan V card has so far is just the greater number of CUDA cores in that product compared to 1080Ti. Same will hold true for Turing 2080Ti I assume.

I still want to see the comparison of compute performance on the 1080Ti and 2080 cards. The change in architecture for asynchronous compute of Turing might have a greater effect on compute performance since the difference in CUDA core count between the 2080 and 1080 is minimal.
9) Message boards : Number crunching : Setting up Linux to crunch CUDA90 and above for Windows users (Message 1956349)
Posted 7 hours ago by Profile Keith Myers Special Project $250 donor
Post:
You are out of luck with a Pascal card as the only way to overclock is to run nvidia-settings.
10) Message boards : Number crunching : GTX is dead long live RTX !! (Message 1956331)
Posted 9 hours ago by Profile Keith Myers Special Project $250 donor
Post:
AnandTech has good scientific benchmarks in their reviews. Hasn't posted yet though.
[Edit] And don't forget Phoronix.com for Linux performance reviews on RTX. Says it should be available later today, IF Nvidia releases any Linux drivers for RTX.
11) Message boards : Number crunching : Intel 16 core/32T discusssion (Message 1956325)
Posted 9 hours ago by Profile Keith Myers Special Project $250 donor
Post:
I would point anyone considering the 24/32C TR2 cpu to go read my earlier post in the thread. Message 1955118
and go view the link I posted by buildzoid. He explains the current requirements of the 24/32C cpus and why only one current board is capable of running them properly and at full capabilities.

The only current board that can do justice to the 24/32C cpus is the MSI MEG Creation X399 board.
12) Message boards : Number crunching : Buying a Supercomputer? (Message 1956270)
Posted 19 hours ago by Profile Keith Myers Special Project $250 donor
Post:
The latest Ubuntu 18.04 LTS distribution has a "minimal installation" option in the installation configuration. It drops the LibreOffice suite, games and all the fluff that the normal full featured installation installs. Just the bare minimum like a editor and browser and basic OS files. That is all that is needed to run the new CUDA90/1/2 special gpu application along with the latest Nvidia driver of course.
13) Message boards : Number crunching : Intel 16 core/32T discusssion (Message 1956267)
Posted 19 hours ago by Profile Keith Myers Special Project $250 donor
Post:
No the TR2 processors are solid. All the benefits of the Ryzen+ advancements. The only questionable products are the new 24 and 32 cores cpus. I would order a 2950 without any concerns. All the benefits of lower latencies, better memory scaling and autoboost that Ryzen+ brought into being. All the 1st generation boards can handle the 16/12 /8 core processors without any concern. Only the new 24/32 core processors exceed the design limits of the 1st gen boards. Only the new MSI MEG Creation board is capable of running those so far. I think that once again AMD has given the board partners little or no advance knowledge that those cpus were imminent. It seems AMD is repeating the same mistake they made on Ryzen launch with processors on the marketplace and nothing to put them into. I don't know who is running the AMD interface teams with the board partners but they have dropped the ball again. I wonder if it was a decision to just launch and steal Intel's thunder first and worry about the boards later.

So, yes I would definitely wait out at least half a year for more TR2 boards to make it to market for the 2970WX and 2990WX chips. I would want more choices than just MSI.
14) Message boards : Number crunching : Intel 16 core/32T discusssion (Message 1956229)
Posted 1 day ago by Profile Keith Myers Special Project $250 donor
Post:
If you want to study up on all the X399 motherboard offerings, a excellent resource is the OCN X399 forums that cover all the X399 boards released so far.
AMD Motherboards
15) Message boards : Number crunching : Intel 16 core/32T discusssion (Message 1956226)
Posted 1 day ago by Profile Keith Myers Special Project $250 donor
Post:
That's a great buy on the 20A Hospital grade outlets. Normal price is $20 a piece for a duplex outlet. The way you can tell a Hospital grade outlet from a common one is the green dot on the face. The outlet is made from nylon and not cheap plastic.

The "best X399" site has the order fairly well developed. The ASUS ZE on the top matches how I would order it with the MSI on the bottom. I too would avoid any refurbed board. No way of telling if any returned board went through a proper manufacturer recertification. The BIOS is best on ASUS, then ASROCK, then Gigabyte and last MSI. That goes for 1st gen boards. Only 2nd gen board so far is the MSI X399 MEG Creation which is top board now because it has the VRM and heat sinks for the 2990WX.

Also pay attention to slot spacing if you intend to run four double wide gpus.
16) Message boards : Number crunching : Intel 16 core/32T discusssion (Message 1955989)
Posted 1 day ago by Profile Keith Myers Special Project $250 donor
Post:
I use hospital grade or at least industrial grade outlets only. The plug clamping force on them is amazingly better than your spec home crap outlets that you find in most houses

Absolutely, Hospital grade outlets. The higher the contact force, the lower the resistance and the less voltage drop and less heating. The material used in hospital grade are better in every way from the metal contacts, the screw connections down to the plastics used. The leakage current allowed for hospital grade outlets are magnitude in difference.
17) Message boards : Number crunching : Intel 16 core/32T discusssion (Message 1955988)
Posted 1 day ago by Profile Keith Myers Special Project $250 donor
Post:
I don't know where you got your chart but it is wrong.

https://www.powerstream.com/Wire_Size.htm

Wrong context. That is for power transmission. As in utility company transmission lines. Also since they are talking about frequency and skin loss, they are talking about RF transmission not 60Hz AC branch circuit used for wiring a house or rooms in a building.
18) Message boards : Number crunching : Intel 16 core/32T discusssion (Message 1955969)
Posted 2 days ago by Profile Keith Myers Special Project $250 donor
Post:
I don't know where you got your chart but it is wrong. 12ga wire is good for 20A. 14ga wire is good for 15A. 10ga wire is good for 30A. Read the chart and notice the double asterisk ** and look for the ** box at the bottom of the page. Copper wire. Derate for aluminum.

National Electrical Code
Allowable Ampacities of Insulated Conductors Rated 0-2000 Volts


And this is the type of cable usually used in American home construction. Normally just called Romex.
Romex® Brand SIMpull®
Indoor Wire Copper NM-B Cable
19) Message boards : Number crunching : Intel 16 core/32T discusssion (Message 1955964)
Posted 2 days ago by Profile Keith Myers Special Project $250 donor
Post:
It's rare to find 240VAC circuits in current U.S. house construction except for one dedicated clothes dryer circuit and maybe a electrical range/stove circuit if a house is all electric and no natural gas installations. All electrical construction was very popular in the States back in the 60's when that design was considered futuristic. If a room hasn't been pre-wired for 240VAC outlets, then the cost to do so is expensive involving new electrical panels, 10ga wiring runs and 240VAC outlets. If your wall/floor/ceiling construction is conducive, it might not be too hard to pull a 240VAC circuit from the panel but it won't be an hours job.
20) Message boards : Number crunching : Intel 16 core/32T discusssion (Message 1955956)
Posted 2 days ago by Profile Keith Myers Special Project $250 donor
Post:
Try and buy any 100% rated breaker at any "big-box" home improvement store. They don't carry anything other than 80% breakers. You would need to special order it or go to a proper electrical supply distributor that carries a full line for contractors.

I had a 20A breaker on that circuit I described. Ran fine at that load for 5 months. Then started tripping regularly. The breaker had been operating at full load for so long that the bimetallic thermal contacts fatigued and started tripping at lower than the 40° C. trip point. Breakers do wear out especially when operated at max capacity. The case that Ian described of thermal runaway was in play. The no load voltage on that circuit is typically about 122-125VAC. I see it down as far as 115-118VAC when operated at full load because of voltage drop and the high current flow in the circuit. Which heats up the wire, outlets and breaker thermal contacts, which increases resistance, which increases current draw, >> repeat cycle.

I investigated all the outlets on that circuit and saw that I had built them with backwired outlets which were legal back in 1986. I thought the NEC had outlawed that by now but when I went to the store to buy proper tag screw terminal outlets, saw backwired outlets still for sale. ??? Guess I need a newer NEC book. Anyway since I knew how much I pull out of the outlets so I spent the $6 for a good quality outlet over the $1 backwired low grade outlets. I still am trying to remember how I wired that room during construction. I thought that each outlet was daisy chained to the next but discovered that isn't the case. I have a star point connection to that room somewhere but I need to crawl under the house and trace how I get to that room from the breaker panel. Could be in a junction box somewhere in the wall too. I can't remember and I didn't document every junction box on the blueprints. I do have each outlet on the blueprint and what circuit breaker number feeds it so that helps.


Next 20


 
©2018 University of California
 
SETI@home and Astropulse are funded by grants from the National Science Foundation, NASA, and donations from SETI@home volunteers. AstroPulse is funded in part by the NSF through grant AST-0307956.