Posts by Alex Storey

log in
21) Message boards : Number crunching : Nvidia GT630 vs 640 vs 650 in SETI? (Message 1482729)
Posted 205 days ago by Profile Alex Storey
Comparing different video cards seems to me not to be a trivial GFLOPs rating comparison nor a whole memory bandwidth matter at all.

If you are talking about the GFLOPS ratings on his Application Details, the problem is it does not show total GFLOPS but GFLOPS per task. He may be running 3 tasks at a time but we have no way of knowing...

If he is running 3, his total would be 3x43.69 GFLOPS but if 2 then 2x43.69 GFLOPS [for SETI@home v7 (anonymous platform, NVIDIA GPU)].

Also the reason Memory Bandwidth is important is when more than one task is crunched per GPU (more Bandwidth helps for more tasks).

But yes, the problem remains. There is no single GPU specification (or even "outside" benchmark for that matter) that hints to S@H performance unfortunately :(
22) Message boards : Number crunching : GPU Wars 2014: Postponed to 2015? (Message 1481393)
Posted 209 days ago by Profile Alex Storey
So the focus really is back on compute! Awesome:)

This paper states this one is stil 28nm, though I suppose could be outdated.

No, no! Not outdated:) TSMC has yet to ship anything 20nm.

SoCs are usually first out the gate and then come the AMD/NVIDIA GPUs. Estimates for the latter are second half of the year, likely late summer.
23) Message boards : Number crunching : GPU Wars 2014: Postponed to 2015? (Message 1481234)
Posted 209 days ago by Profile Alex Storey
With my relatively new (and self-proclaimed) role as Cassandra of these boards I'm going to go ahead and guess that Boinc is just "seeing" the flops wrong (it's not like it runs any benchmarks to get that number, just blurts out whatever it's told AFAIK).

The Gflops for the 448 are spot on but the Gflops for the 750ti should be around 1300 (according to Wiki anyway, which can be a bit dodgy of course).

I do hope I'm wrong though.
24) Message boards : Number crunching : GPU Wars 2014: Postponed to 2015? (Message 1481100)
Posted 210 days ago by Profile Alex Storey
I know it's early days still but how does it compare against the 560Ti?

It seems to be 2x faster !

If that's true then it's the biggest news since the move to the colo.

I was going to say I'm surprised the forums haven't caught fire but then the cynic in me tapped my shoulder and promptly reminded me that there is hardly ever any number crunching done in the Number Crunching forum:)

But are u sure? Because before you posted I had a look at your tasks and for some reason I came to the conclusion that the 750ti can do about 80% of what the 560ti can do. Specs would be closer to these numbers too but... specs and Seti never seem to go hand in hand.

Hopefully I'm wrong and my guess on how many tasks you were crunching on each card was off:)
25) Message boards : Number crunching : GPU Wars 2014: Postponed to 2015? (Message 1480712)
Posted 211 days ago by Profile Alex Storey
I know it's early days still but how does it compare against the 560Ti?
26) Message boards : Number crunching : GPU Wars 2014: Postponed to 2015? (Message 1480215)
Posted 212 days ago by Profile Alex Storey
Hi, I just took delivery of 3 x 750 Ti and will fire them up in a couple of hours.
I will post back here if they work or not.

Any suggestions as to which version I should run Cuda 42 or Cuda 50 ?


CUDA 50 would be your best bet.

I'm not pretending to be even remotely qualified to take a shot at answering but I do want to say I wouldn't be surprised if it turns out to be Cuda 42. As I understood the Anand article, Maxwell architecture is closer to Fermi than Kepler. If that is truly the case and I'm not oversimplifying my reasoning (though I probably am) that could even mean that 42 works while 50 doesn't? We'll know soon enough:)

Congratz on your new cards RED and if you are up for it, you may need to consult with someone like Richard or Joe (or anyone who thoroughly understands Server-side validation) in case it might be a good idea to run stock for a few days or until you complete 10 tasks from each app? I honestly have no idea if this would help the project in any way, just bringing it up so a guru can shoot it down or expand on it:)

Edit: Apologies for getting your name wrong Reg :)
27) Message boards : Number crunching : Rescheduling - final attempt (Message 1480132)
Posted 212 days ago by Profile Alex Storey
...i canĀ“t agree with you... you insist in point someone directly, that only could flame the thread.

I was about to say the same. Although maybe not nice enough to pass the "annoy or antagonize other people" & "messages that are deliberately hostile or insulting" rules of moderation :P

On a side note, it has made sense for years to build many & mediocre crunchers for Seti. I've said it before & I'll say it again: my netbook has never starved for work in over 3 years, except for when the Lab shut down for a week or more. A smart person would build as many single-card (let's say) GT 650 machines as they could afford if they really wanted to avoid the stress & babysitting.

I think David Anderson called it "the hot-rod mentality" in an Nvidia keynote, many years ago. Now I'm all for "hot-rods"... I really am.. they just haven't made sense for years, as far as Seti is concerned.

On another side note, these threads are just a prelude. I imagine they will get worse before they get better. It appears Raistmer has kicked a hornets nest and that is almost definitely a good thing. What he is doing manually now will likely be done automatically (based on his experience & findings) one day by Boinc. Sure it'll take like 5 years but whatever:)

On another-other side note, going up against a volunteer guru has the painfully obvious effect of biting the hand that feeds you. And it's OK to bite the hand that feeds you but you better have a faaaaar better argument than dogma.

Edit: And this from someone running stock apps and AP de-selected:)
28) Message boards : Number crunching : GPU Wars 2014: Postponed to 2015? (Message 1479762)
Posted 213 days ago by Profile Alex Storey
A few random tidbits:

@Tullio - OC stands for (is the abbreviation of) "overclocked", so yes, you were looking at an overclocked card... At the bottom of this link is a chart and - in bold - has the differences between the 750 and 750 OC. A few columns over is the 750Ti too.

@archae86 - always nice to see you drop by :)

Forgot to mention that since this is a new architecture, there is a chance it may not work out-of-the-box with Seti.
29) Message boards : Number crunching : Rescheduling - final attempt (Message 1479746)
Posted 213 days ago by Profile Alex Storey
I wanted to post this in Raistmer's thread but then it got locked so I sent him this PM yesterday (Now that this thread has opened up I'm posting it here)

Since you are making a lot of sense, I had a quick think to figure out why no-one understands what you are saying. I may have found one of the reasons....

Boinc inability to enforce limits on a per-app basis. You know this well but it hasn't been mentioned? (Didn't follow Juan's thread)

SETI@home v7 7.00 windows_intelx86 : 50 tasks
SETI@home v7 7.00 windows_intelx86 (cuda23): 200 tasks
AstroPulse v6 6.04 windows_intelx86 (opencl_nvidia_100):100 tasks

Of course I just made the numbers up. They are not important. The fact that they should be treated differently is what is important. You and Jason (I think) have complained about this before but it was probably over on Beta... Maybe you should explain it here?

Sorry if I misunderstood something!
30) Message boards : Number crunching : GPU Wars 2014: Postponed to 2015? (Message 1478872)
Posted 215 days ago by Profile Alex Storey
Haven't had time to digest these but at first glance it looks like we're back to the good ol' days of Fermi (sans toaster jokes!) :) I think Kepler proved hard to get excited about as far as Seti is concerned. Maxwell may end up far more promising but as we all know by now, specs alone can be deceptive when it comes to Seti...

Here are your links for today:

AnanadTech - Compute (page 21 of 24)
The NVIDIA GeForce GTX 750 Ti and GTX 750 Review: Maxwell Makes Its Move

TPU - Performance per Watt (page 26 of 30)
NVIDIA GeForce GTX 750 Ti 2 GB
31) Message boards : Number crunching : ASUS ROG MARS760-4GD5 GeForce GTX 760x2 (Message 1475852)
Posted 222 days ago by Profile Alex Storey
Furmark Power consumption 402 W !! :)

Edit: just for the card, not the whole system
32) Message boards : Number crunching : Android x86 app (Message 1472327)
Posted 230 days ago by Profile Alex Storey
Any news on those results?

Sorry, couldn't make it to the forums any sooner...

Unfortunately, things don't appear to be running smoothly (for me) out of the box. You can find more info in the x86-android thread, over on the Seti Beta Forums:
33) Message boards : Number crunching : Happy New Year!!! (Message 1459525)
Posted 263 days ago by Profile Alex Storey
Happy 2014 from Greece :)
34) Message boards : Number crunching : Observation of CreditNew Impact (4) (Message 1444501)
Posted 305 days ago by Profile Alex Storey
At first glance it appears CN is auto-adjusting every few tasks. Assuming all tasks want to settle around 38, it looks like it's trying to compensate whenever you get low-balled and vice-versa. Which could also be an explanation for repeating patterns?

Why it has to jump through these hoops is, of course, a different kettle of fish. But it looks (dare I say) fair?
35) Message boards : Number crunching : Observation of CreditNew Impact (4) (Message 1444236)
Posted 306 days ago by Profile Alex Storey
OK so while it's unfortunate that 'Boinc whetstone' is out of whack it has little to do (maybe AVX portions?) with the drop in credit, correct?
(Please understand I realize how important this discovery of yours is, I'm just asking if it has any relevancy to the credit drop)

So let's focus on 'undamped weighted averages'. I'm guessing that 'stable continuous control loops' were somehow vulnerable to some kind of cheating and therefor replaced by 'fudged sampled averages'?:)

Next question will be 'what exactly are undamped weighted averages?' but I'll need to brush up on the CN wiki to ask it well. So maybe tomorrow if u are still in a question-answering mood:)
36) Message boards : Number crunching : Observation of CreditNew Impact (4) (Message 1444224)
Posted 306 days ago by Profile Alex Storey
Ask me any questions on symptoms/observations you like. I will do my best to respond at your required tech level.

Jackpot! Q&A with a Seti guru:) Thank you Jason.

I see you quoted my pet theory but until I re-read the CN page and put pen to paper I can't really form a proper question to ask. So here's a couple of other questions:

a) This whetstone discovery of yours applies to V6 too right? Does it go all the way back to Eric's flop-counter? Before the days of CreditNew?

b) And probably the hardest and most obvious question to ask: Why did credit drop? I've read replies from you, Claggy, Joe, Eric, and others but I still can't get my head 'round it.
37) Message boards : Number crunching : Observation of CreditNew Impact (4) (Message 1444149)
Posted 306 days ago by Profile Alex Storey
It would be very interesting if any of the posters complaining of reduced RAC were running stock apps: - Anybody here for whom that applies?

Yeah, I ran stock V6 for a little over a year... and then optimized V6 and now stock V7. My numbers (RAC) when crunching 24/7 were:

stock v7: 1000
stock v6: 1500
opti V6: 2000

Here's a picture of my opti V6 changeover to strictly V7 stock:

But Alan, I don't think the answer to your question is that simple. My laptop has a GPU while yours does not. For a CPU only machine, I'm guessing the transition from stock V6 to stock V7 would actually show a slight increase in RAC! This can easily be seen by your own APR numbers from your host:

To complicate things even further the stock V6 GPU app was slightly faster than the optimized V7 GPU app. In an ironic reversal of roles & goals the stock app was optimized for speed while the anonymous app was optimized for better science. I guess this makes sense considering Nvidia made the stock app while Seti gurus & project-lovers made the opti app:)

However the optimized V6 CPU app was nearly twice as fast as the stock CPU app. Which is why I still have the nagging feeling that the servers were actually using the opti V6 CPU app as a benchmark rather than the stock V6 CPU app.
38) Message boards : Number crunching : Observation of CreditNew Impact (4) (Message 1444046)
Posted 306 days ago by Profile Alex Storey
In short, v7 WUs generate circa 69% of the credit that AP WUs generate per day. Jason, this is why your comment in the other thread is not right.

Apples and oranges:)

I think your misunderstanding lies in that you took Jason's comment to be v7 exclusive. The way I understand it, it applies to v6 too (though maybe with different percentages for v6 MB but that's irrelevant here). In other words it has nothing to do with your quest of why credit was halved going from v6 to v7. Rather it's a new discovery of a problem that dates back to 2009 (?) and not this summer.

I think where you got confused is that he indirectly answers your question of why MB vs. AP is showing a discrepancy in credit (answer: because they are calculated separately). But Jason's numbers have nothing to do with any numbers you are witnessing "today".

So I guess what he was trying to say was that the numbers aren't as bad as you think they are. They're worse. So bad, in fact, that any attempt to micromanage your crunching is futile (philosophically speaking).

On a side note, I think a lot of people here are underestimating precisely how intelligent CreditNew really is. Any attempts to tweak credit-granting (the ones you are calling "less than satisfactory" or maybe even "half-hearted"; I disagree) have been slapped down by CN. This can only mean they were anticipated (a priori) by design. A fact which I find fascinating, unexpected and very, very clever.
39) Message boards : Number crunching : Costs per Credit? (Message 1443697)
Posted 307 days ago by Profile Alex Storey
I'm not exactly qualified to answer but I'll have a go 'till someone else comes along:)

Don't know if it really is the audio drivers causing your crashes but since I can't imagine someone using 2x760s as an HTPC, you probably don't need to install those NVIDIA (audio) HDMI drivers.

I can't say I remember the procedure that well but when doing a manual/clean-install (which you should always do) of NVIDIA drivers I think you get 4 boxes of stuff to choose from. The actual graphics drivers, which are the only thing you really need, are greyed-out and will install no matter what (so technically that leaves 3 boxes to choose from).

All the other stuff is optional. HDMI you probably don't need and same goes for PhysX (unless you play games that actually use it; in which case you would, of course, install it). I think there is a 3rd tick-box too but can't remember what it is, I just remember that I never check (as in I always un-tick) all of the optional boxes.

Hope this helps
40) Message boards : Number crunching : Haswell Temps??? (Message 1442616)
Posted 310 days ago by Profile Alex Storey
I didn't have a look at Lionel's link (I'm sure they say everything I'm about to say) but general consensus (I have no personal experience with Haswell) from articles and comments around the web is that this is 'normal'. Haswell was a huge disappointment for enthusiasts unfortunately.

You'll have to search this 'cause I don't remember well but I think Intel cheaped out on the way they cover the chip*. Instead of solder they're using a fancy glue? I seem to remember some crazies stripping the casing and cooling the actual chip directly...

Edit: Another link
Intel Core i7-4770K CPU Review. Intel Haswell for Desktops: Ruin of Our Hopes?
(page 11 of a 13-page article)

*this may have happened/started on Ivy Bridge?

Previous 20 · Next 20

Copyright © 2014 University of California