5748 WU

[sirmonkey1985]

Pretty much the same as everyone else. WU 5748 dropped about 2000 Ppd and temp up about 3 degrees C on EVGA 8800GS. No other problems. G

woohoo still havent gotten the 5748 core... go me and the 1.15 core

 

[Xilikon]

Vincent Voelz, who is in charge of this project, said that it's now public but with a low assignment weight so not everyone will get them yet until they raise the assignment weight.

 

[alan2308]

n/m beaten to it

 

[Pooch666]

woohoo still havent gotten the 5748 core... go me and the 1.15 core

You'll get one sometime and it will automatically install 1.19 core. G

 

[SatTech]

Checking the stats page it seems I've done 4 of these. No wonder my production was 'off' yesterday.

 

[jws2346]

Undaunted by minor PpD setbacks, here are my PpD results according to FAHmon v2.3.4. They have been rounded off because I'm basically lazy. I have two other GPU's. They are a "single slot" PNY 8800GT and a "dual slot" eVGA 9600GSO. I don't have any statistics for the "big bad" 5748 WU on either GPU card yet. (key word there is yet )

eVGA 8800GS, shaders 1601, fan manual 80%, core temp 53c (EVGA Precision), ambient temp 24c
WU 5748 = avg time = 2mn 43s 2700 PpD
WU 5606 = avg time = 1mn 40s 4100 PpD
WU 5015 = avg time = 1mn 34s 4400 PPd

Arrgh..., the point reduction is kind of a "bummer", but we all knew it was coming (one thing I've read many times, F@H will always change ). Looking at it on the "bright and altristic side" (the most important side) we're doing a project to help scientists and medical professionals find cures for some super fu*ked up diseases (Alzheimer's, Mad Cow (BSE), CJD, ALS, Huntington's, Parkinson's disease, and many Cancers) . IMHO the world's people would be so much better off if the unfortunate victims of those super fu*cked up diseases (Alzheimer's, Mad Cow (BSE), CJD, ALS, Huntington's, Parkinson's disease, and many Cancers) could be cured or at least not have to suffer as much. You gots to use current tech. Just think you'd still be digging holes to poop in (or getting splinters in your butt) and going to the river for water or bearting the shite out of a horse to get your frozen TV dinners to the creek before they thawed or croaking from no penicillin (like dyin' from the "clap") without current technology. ( no refrigerators to use the ice in the winter or something like that, you mean "Wild Bill" didn't have TV dinners, holy shite, no microwave either )

I've gots to stop, I'm depressing myself. Bring on the biggest, baddest WU's you gots Stanford. The [H]orde will crunch the shite outa' them. (please just make sure the WU's and clients are stable)

Folding and WCGing for the CURE

 

[sirmonkey1985]

You'll get one sometime and it will automatically install 1.19 core. G

good thing i have a backup of the 1.15 core..

 

[AssM@N]

No ones gonna take the hit in PPD as hard as I am..... my farm is 100% pure GPU based!

O well....gonna take much longer to catch FLECOM now

ROC

 

[APOLLO]

No ones gonna take the hit in PPD as hard as I am..... my farm is 100% pure GPU based!

O well....gonna take much longer to catch FLECOM now

ROC

Yeah, 8 months instead of 5 and a half....

 

[sirmonkey1985]

finally got one.. not to bad.. ppd went to 3800.. but only takes 3 hours instead of 2 1/2 hours.. im not complaining..

 

[Kowan]

My GTX280s finish a non 5748 in 1h 20m and FahMod shows a PPD of 8640 to 8823.
The 5748s take 1h 56m and show 6589 PPD.

I can hear when one is downloaded. The card whines for help.

 

[sirmonkey1985]

My GTX280s finish a non 5748 in 1h 20m and FahMod shows a PPD of 8640 to 8823.
The 5748s take 1h 56m and show 6589 PPD.

I can hear when one is downloaded. The card whines for help.

so far i havnt had a problem.. still running the same clocks as before.. but the one thing i did notice that it uses only 1% cpu unlike before it was using 11% even with wcg using both cores..

 

[osrk]

Bah got hit with the work unit tonight on one of my GX2's, had to shut down one core because the temp sky rocketed over 10C, only running three cores for now until I can run to radioshack and get a PCI slot cooler. Haven't gotten hit with that work unit of the other GX2 or my ATI cards.

 

[sirmonkey1985]

Bah got hit with the work unit tonight on one of my GX2's, had to shut down one core because the temp sky rocketed over 10C, only running three cores for now until I can run to radioshack and get a PCI slot cooler. Haven't gotten hit with that work unit of the other GX2 or my ATI cards.

yeah i just noticed my gpu temp went from 45C to 49C but im running a duorb cooler in an antec 300 case with 5 fans not counting the psu fan right below the card..

 

[Hgradio]

As far as points go, for us GPU peeps it is like a football field, The turf is the same for both teams, muddy or dry, lol.

 

[AMD_Gamer]

I have gotten a few of these so far today

Do these WU's actually do more work or what? how exactly does it work or is the PPD just messed up like they assign a low point value for them?

I'm all about the WU's and not the points but i rest better at night knowing that i am actually folding a WU that is bigger and more scientific info and not just a point error?

 

[sirmonkey1985]

I have gotten a few of these so far today

Do these WU's actually do more work or what? how exactly does it work or is the PPD just messed up like they assign a low point value for them?

I'm all about the WU's and not the points but i rest better at night knowing that i am actually folding a WU that is bigger and more scientific info and not just a point error?

http://fah-web.stanford.edu/cgi-bin/fahproject?p=5748

 

[Xilikon]

I have gotten a few of these so far today

Do these WU's actually do more work or what? how exactly does it work or is the PPD just messed up like they assign a low point value for them?

I'm all about the WU's and not the points but i rest better at night knowing that i am actually folding a WU that is bigger and more scientific info and not just a point error?

The points value is correct. I already warned people way back in august about those "Lambda" units and the inherent points hit. The reason is because units is benchmarked on a ATI card so if ATI perform better with them, NVIDIA will be worth less than this. However, the scientific value is very great and the reason the Pande Group kicked the 1.09 core in the bucket even if it's stable is because it's not stable with those Lambda units.

The PG mentionned that eventually, all the units will be Lambda so we are gonna learn to live with them. The high PPD is a "bonus" we got since the start and those bonuses are not eternal. As Hgradio said, everyone will suffer equally

 

[osrk]

The points value is correct. I already warned people way back in august about those "Lambda" units and the inherent points hit. The reason is because units is benchmarked on a ATI card so if ATI perform better with them, NVIDIA will be worth less than this. However, the scientific value is very great and the reason the Pande Group kicked the 1.09 core in the bucket even if it's stable is because it's not stable with those Lambda units.

The PG mentionned that eventually, all the units will be Lambda so we are gonna learn to live with them. The high PPD is a "bonus" we got since the start and those bonuses are not eternal. As Hgradio said, everyone will suffer equally

Oh this is so mean of them dropping our point values!! I'll show them time to order a few more ATI 4850's! (*wink*)

Hehehe they can run whatever they want i really don't care about point values. i use them to make sure each client is running at it's max potential.ie if gpu 1 is crunching 5k ppd then gpu 2 should crunch 5k ppd with the same WU.

 

[sl1982]

Hmm the new wu cause eue on my gpu. Even when underclocked. Seems the little gpu that could finally hit a brick wall.

 

[psyshack]

Ive ran them on all of my gpu's. No issues yet except for the massive points hit.

 

[Wheresatom]

I just noticed my first one. I am not sure if it is a coincidence or not, but I also got a BSOD. It is hot in my bedroom, so it may be unrelated, but I am going to scale back a notch on my OC in the mean time.

 

[Kendrak]

I just noticed my first one. I am not sure if it is a coincidence or not, but I also got a BSOD. It is hot in my bedroom, so it may be unrelated, but I am going to scale back a notch on my OC in the mean time.

I bet the temps when up, Vid card stopped responding and BSOD.

I've had to when I forget to turn off folding and load up a game

 

[EvilAlchemist]

The points value is correct. I already warned people way back in august about those "Lambda" units and the inherent points hit. The reason is because units is benchmarked on a ATI card so if ATI perform better with them, NVIDIA will be worth less than this. However, the scientific value is very great and the reason the Pande Group kicked the 1.09 core in the bucket even if it's stable is because it's not stable with those Lambda units.

The PG mentionned that eventually, all the units will be Lambda so we are gonna learn to live with them. The high PPD is a "bonus" we got since the start and those bonuses are not eternal. As Hgradio said, everyone will suffer equally

X you are 100% correct. We have been saying this for a very long time.

Oh this is so mean of them dropping our point values!! I'll show them time to order a few more ATI 4850's! (*wink*)

LOL .. I already have 10 HIS 4850's running ....

BTY: An HIS 4850 ICEQ version uses about 25% less power then a stock ATI 4850.
Stock 4850 = 110w max ICEQ / IceQ Turbo = 85w max
(For those of you who like to know that stuff)

 

[jebo_4jc]

^^^ wow how did HIS pull that off?

 

[APOLLO]

X you are 100% correct. We have been saying this for a very long time. LOL .. I already have 10 HIS 4850's running ....

Wow, I think you set a record for Ati cards on this team.

BTY: An HIS 4850 ICEQ version uses about 25% less power then a stock ATI 4850. Stock 4850 = 110w max ICEQ / IceQ Turbo = 85w max
(For those of you who like to know that stuff)

Thanks for the info. I prefer HIS to other Ati brands. The only Ati card I have running right now is a Sapphire only because I couldn't find HIS locally.

 

[Tigerbiten]

Posted by VijayPande over on the Folding Forum ............

Here's the scoop for why the PPD may be lower. It's a combination of two effects. First, the NV core does really well compared to ATI for small proteins. However, we're moving to larger proteins and there NV has less of an advantage. The benchmarking is still the same (and if you have any doubts, you can compare performance of similar projects on ATI hardware similar to our benchmark machine). The second effect is that we rolled out NV core v1.19 to improve stability. This new core is slower but (for most) much more stable.

The combination of a core which is ~10-20% slower and bigger proteins have brought down the PPD. However, note we've discussed this before (in fact the last time we rolled out bigger proteins). In that case, we found immediately that the bigger proteins were having trouble with the core at the time (1.09) and that forced us to make some code improvements for precision (new code for the integrator and constraints) and that's been rolled into v1.19. It looks like the EUE rate is way down, which is great, and we're making good science progress.

One final comment: NV core_13 and core_11 are now in sync. This project was released with core_13 as a test, but then was moved through QA and is now a production project. To avoid confusion, it should have been rolled out as a core_11 project. That wouldn't change the science (and in fact would set us back a bit while we QA an identical project), but probably would have been good to make it more clear. I've talked to the scientists involved with this project about this and now this is more clear to them.

So to summarize, this is not a misassignment, in the sense that this is a production project. However, it would have been more clear if this came out as core_11 in terms of people's understanding of what's going on. I've also stressed to the scientists involved that they should make a post to make this more clear as they roll out projects in the future.

So it look like the lower points are here to stay ............

Luck ..............

 

[Razor_FX_II]

Of coarse it had to be the Tiger to bite me.
All Nvidia here. Looks like it's time for more hardware.

 

[EvilAlchemist]

^^^ wow how did HIS pull that off?

The HIS ICEQ cards are a non-refrence design. I don't know how they did it, but it explains why my wattage on my systems are so low.

So it look like the lower points are here to stay ............

For nVidia .... ATI --->

All the gpu will still fold, and the PPD may go back up after the next core optimizations go back in on the nVidia side. ATI clients have been going up in PPD all the time.
Either is a good choice ... but that is why i kept going with ATI.
I had a feeling this would happen.

And yes, I think i am the ATI F@H person here now

 

[psyshack]

So how are the ATI cards folding on the big proteins compared the the whimpy NV cards?

Are the ATI cards blowing away NV? Are they running the same points yet on the same or same family of proteins?

I keep reading how the ATI cards are so great and a few bucks cheaper. But cheap doesn't pay AMD's bills.

So how are these monster cards fairing in the large protein world?

 

[EvilAlchemist]

So how are the ATI cards folding on the big proteins compared the the whimpy NV cards?

Are the ATI cards blowing away NV? Are they running the same points yet on the same or same family of proteins?

I keep reading how the ATI cards are so great and a few bucks cheaper. But cheap doesn't pay AMD's bills.

So how are these monster cards fairing in the large protein world?

ATI is almost the reverse of nVidia.
ATI cards seem to preform very well, the larger the protein gets.
nVidia cards are really fast on small proteins, but slow on larger ones.

ATI & nVidia work units are generally the same.
Any project can be configured to run on any GPU, but this is done at the Stanford End.

You also have to remember that ATI & nVidia have differenct marketing goals.

ATI approach is to have a good gpu, but add multiple for better performance.

nVidia goes it to make very large and fast GPUs.

GPU makers don't make money on the high $ items, they make the bulk of their sales to OEM's of mid and low end GPU's.

CPU Mag had a great article about this a few months ago.
I will try and post a copy of it here, for you to read up on ...it was a very good read.

 

[jws2346]

Yeah, come on Evil, share some of that ATI credibility and perserverance (sticking with ATI even when they were "doggin' on the PpD scene). Do the ATI cards, with the new, improved, bigger WU's give more PpD for a ATI card as compared to nVIDIA cards. (in other words are the nVIDIA cards the dog bait now, only kiddin')

Bad news bears, I had to retire my 9600GSO boxen for a few, it blew a HD. I ordered another one (it even blew an oldy 10 gigger I had stashed in reserve for Linux) I'm just waiting for the "pony express" to deliver it. (UPS)

Overheatin' transisters Folding and WCGing for a CURE

 

[psyshack]

ATI is almost the reverse of nVidia.
ATI cards seem to preform very well, the larger the protein gets.
nVidia cards are really fast on small proteins, but slow on larger ones.

ATI & nVidia work units are generally the same.
Any project can be configured to run on any GPU, but this is done at the Stanford End.

You also have to remember that ATI & nVidia have differenct marketing goals.

ATI approach is to have a good gpu, but add multiple for better performance.

nVidia goes it to make very large and fast GPUs.

GPU makers don't make money on the high $ items, they make the bulk of their sales to OEM's of mid and low end GPU's.

CPU Mag had a great article about this a few months ago.
I will try and post a copy of it here, for you to read up on ...it was a very good read.

Still does not answer the question. Are they still in the tune up stage? Have they come on strong yet with the larger proteins?

I need a excuse to go buy a ATI card. If the tables are going to turn in the near future,,, great! But if its still,,, There coming, the drivers are getting better or any number of reused excuse's out of the ATI camp then I will set on my cash.

I just figured with the bigger proteins coming out there would be something of credit possible out there now to support the claims.

Not hating so please do not take it that way.

 

[EvilAlchemist]

Still does not answer the question. Are they still in the tune up stage? Have they come on strong yet with the larger proteins? .

ATI has had 0 large / supersized proteins in recent months.
There were some very early on and the points per day were great.
You have to remember that was on a older core, so no telling how good the points will be.
If I knew a number, I would give it too you, but I just dont ...

They are working on the ATI core all the time, and there is a lot of room for speed improvments ... just going to take time to get there and test them all out.

I need a excuse to go buy a ATI card. If the tables are going to turn in the near future,,, great!

I can't warn you enough not to buy hardware on what the current PPD structure is.
First we all purchased Q6600 cause of SMP .. then we all wanted nVidia GPU's ... now ...

It is a never ending cycle .....
I like my ATI GPUs (Specificially the ICEQ versions)
They are stable and work for me ....

Folding@Home is like Vegas ... go there expecting to lose it all.
If you win great .. if you lose ... enjoy the game and have fun ...

I just figured with the bigger proteins coming out there would be something of credit possible out there now to support the claims.

I really wish I had some data to give you. I could make a guess at what I *think* it will be, but it would just be a guess. I think ATI cards will do well in the up coming work units.
I truly think that would would not regret getting an ATI card. No need to buy 5 or 10, but get one and let it run.

 

[Kendrak]

It is a never ending cycle .....
I like my ATI GPUs (Specificially the ICEQ versions)
They are stable and work for me ....

Folding@Home is like Vegas ... go there expecting to lose it all.
If you win great .. if you lose ... enjoy the game and have fun ...

Just don't lose it in a puff of white smoke..... magic show or not.

 

[alan2308]

Just don't lose it in a puff of white smoke..... magic show or not.

You haven't truly earned your [H] until you've lost something in a violent and spectacular fashion.

 

[Kendrak]

You haven't truly earned your [H] until you've lost something in a violent and spectacular fashion.

This was a few years ago..... but does filling a room whit white smoke from a monitor work