One real wild card in any such decisions is the unknown matter of how GPU-using applications available next year on Einstein will behave, both in terms of GPU hardware fit and in terms of host requirements. The recent cuda55 variants of both the BRP6 and BRP4G applications have had pretty modest host impact, but there is nothing fundamental about that. There is reason to suspect a hoped-for GW application using GPU power may have much higher relative CPU demand, for example.
It would be useful if they could reduce the GPU memory bandwidth requirements, so that more cards (whether Maxwell or Pascal) can live up to the potential exemplified by the GTX 750 Ti. I am sure the Einstein developers are well aware of that, but whether they can do anything about it remains to be seen. I would suspect that some of the new apps will require significantly more bandwidth than others, and we will need to match cards to apps accordingly for best efficiency.
The next members of the Pascal family to ship apparently will be the 1050 and 1050Ti, within days.
As I have believed for quite a while that the 750/750Ti cards are remarkably good Einstein cards, with fabulous power efficiency for their generation, good Einstein performance per unit price, and easy fit into systems, I look forward to seeing actual Einstein results with these new cards.
The announced stock clock rates are pretty low compared to other Pascals. It remains to be seen what the practical overclocking limits for memory and core will turn out to be, running Einstein. As the memory bus is only 128 bits wide, these certainly are going to be far lower in Einstein performance (at least on the recently retired BRP6 application--which I understand better than others) than the higher grade Pascal cards, but the rumored prices are very low, with launch prices of $109 for the 1050, and $139 for the 1050Ti mentioned.
Regarding power, although both cards state a 75W TDP, they both lack a supplementary power connector, and in general the Pascal cards have been burning considerably less than TDP while running Einstein work. It remains to be seen whether these are remotely so good a power outlier within line as were the 750s, but at least they seem likely low enough as to pose little problem of existing supply adequacy for most people considering an upgrade.
For before and after launch discussion, and links to reviews when they come out, one source is this reddit thread.
While the users who built four 980Ti cards into one system may not find these of any interest, I think people with older hardware toying with the idea of a graphics card upgrade who have some sensitivity to purchase cost and electric power operating cost may find them pretty interesting.
One real wild card in any such decisions is the unknown matter of how GPU-using applications available next year on Einstein will behave, both in terms of GPU hardware fit and in terms of host requirements. The recent cuda55 variants of both the BRP6 and BRP4G applications have had pretty modest host impact, but there is nothing fundamental about that. There is reason to suspect a hoped-for GW application using GPU power may have much higher relative CPU demand, for example.
I like the 1050Ti, I will buy one when they release. Currently I run six 750Tis and enjoy their low power efficiency.
Videocardz has a web page listing a great many specific manufacturer part numbers for 1050 and 1050Ti cards, and states the expected release date as October 25, 2016.
Some companies have web pages up for their cards. For example this PNY base model 1050 stated to list for $119.99. If connectors matter to you, don't assume the 1050s will follow the Pascal trend. This one has only one DisplayPort, one HDMI, plus one "DL-DVI". As the bigger Pascals I own lack DVI entirely, this may be good news to some.
Another point I missed initially on the reddit thread I earlier linked is that all of these 1050 chips are stated to be manufactured on the Samsung 14nm FinFET process. This differs from the TSMC processing for the bigger Pascal chips, and may well shift power vs. frequency characteristics, and practical maximum operating frequencies, possibly considerably. This makes guessing Einstein performance based on simple ratios to known Pascal cards more speculative than I like.
Another point I missed initially on the reddit thread I earlier linked is that all of these 1050 chips are stated to be manufactured on the Samsung 14nm FinFET process. This differs from the TSMC processing for the bigger Pascal chips, and may well shift power vs. frequency characteristics, and practical maximum operating frequencies, possibly considerably.
That is quite interesting. So it is a GP107, and not a cut-down GP106. It could have its uses.
I'm really hoping the GTX 1050 Ti has the same layout for core and memory as the higher end cards, because it would be nice to be able to pick up an Asus Strix block and waterblock a few of these lower end cards. Why lower end ones instead of the higher end ones? I want two cards in a rig, but can only use up so much power, and want to put a pair of them in alongside with a dual E5-2670 rig ^_^
I think the review embargo released some hours ago, and there are quite a few reviews posted. As most of them are quite firmly targeted at game-players, squeezing out information relevant to running Einstein is a bit iffy.
Some general points:
1. quite low power (at least one reviewer specifically called out slightly lower power than a GTX 750 at game settings)
2. extremely low power at idle (good if you don't BOINC 24/7, but I do, so this matters little to me).
3. Some greatly prefer to the 1050Ti to the base 1050, others say that, especially because clock rates come set faster on the 1050, that the extra performance of the 1050Ti is not likely worth the extra price and thus greatly prefer the 1050.
4. That last argument may crucially depend on how 4 vs. 2 Gbytes of VRAM actually affects you.
5. a dual-fan version for which I saw a review is really, really quiet.
6. a lot of the models of the cards are not widely available at all yet. In a quick look I did not spot any on offer at Amazon, and Nowinstock had not yet started a tracker.
7. for those of us who overclock, it is disconcerting that some hard upper limits on both memory and core clock speed were observed. As Einstein work is not likely to exceed the power limits, these artificial rate limits may stop us from getting more out of the cards. On the plus side, maybe fewer of our participants will run rigs returning high error fractions because of excess zeal and inadequate checking and monitoring.
I'm currently inclined to get one, but am not at all settled on which one. I'm pretty fan noise allergic (and have paid many extra pennies on more expensive case fans), so should probably get a dual-fan model. I currently have two 750Ti and one base-model 750 running, and would most likely do a direct swap, allowing comparison on the chosen platform.
I ordered an EVGA 1050Ti SC single fan model. All of my current 750Tis SCs are mostly single fan and run the same or cooler then my dual fan FTW 750Ti for similar clock speeds. I have one lil 750Ti that has an ASIC value of 89.9% and boosts almost 1400mhz on its own. Will the 1050Ti be the same? hope so since I like the simplicity of not needing the six pin connector. If I was going to OC the card further I would look into the dual fan ACX 3.0 version EVGA sells. I'm expecting a 20-30% increase in productivity with this card vs my exsisting 750s. I'm also a big fan that it comes with 7000mhz 4GB ram.
The whole announcement availability structure seems a bit different for the 1050 than the earlier Pascal cards. As of now (a couple of days after release day) I still don't spot any listings on Amazon (not even the third party very high priced, buy now get shipped later stuff). NewEgg, on the other hand has several listings, and among the 1050Ti cards listed, many actually show as available for shipment. BH photo also has a few listings, but only for backorder.
My current intention is, mostly for curiosity, to buy a dual-fan 1050 (not Ti), prefering one that is short, cheap, and likely to have low and agreeable fan noise. I've backordered a Gigabyte GV-N1050OC-2GD, but that may well change if something I like better turns up before that ships.
By the way, I think people putting one of these as the only graphics card in a moderately well-ventilated case who are not unreasonably bothered by fan noise can safely get single-fan models. But for my usage none of these three conditions are met, so I choose to limit my choices or spend a little more to get two fans.
Memory OCs help production quite a bit with E@H. With the reduced mem speed and width I don't think these cards will scale down just based off the # of cores. I can easily push my 1070 to +900 MHz running E@H and every bit of it helps.
archae86 wrote:One real wild
)
It would be useful if they could reduce the GPU memory bandwidth requirements, so that more cards (whether Maxwell or Pascal) can live up to the potential exemplified by the GTX 750 Ti. I am sure the Einstein developers are well aware of that, but whether they can do anything about it remains to be seen. I would suspect that some of the new apps will require significantly more bandwidth than others, and we will need to match cards to apps accordingly for best efficiency.
archae86 wrote:The next
)
I like the 1050Ti, I will buy one when they release. Currently I run six 750Tis and enjoy their low power efficiency.
Videocardz has a web page
)
Videocardz has a web page listing a great many specific manufacturer part numbers for 1050 and 1050Ti cards, and states the expected release date as October 25, 2016.
Some companies have web pages up for their cards. For example this PNY base model 1050 stated to list for $119.99. If connectors matter to you, don't assume the 1050s will follow the Pascal trend. This one has only one DisplayPort, one HDMI, plus one "DL-DVI". As the bigger Pascals I own lack DVI entirely, this may be good news to some.
Another point I missed initially on the reddit thread I earlier linked is that all of these 1050 chips are stated to be manufactured on the Samsung 14nm FinFET process. This differs from the TSMC processing for the bigger Pascal chips, and may well shift power vs. frequency characteristics, and practical maximum operating frequencies, possibly considerably. This makes guessing Einstein performance based on simple ratios to known Pascal cards more speculative than I like.
archae86 wrote:Another point
)
That is quite interesting. So it is a GP107, and not a cut-down GP106. It could have its uses.
I'm really hoping the GTX
)
I'm really hoping the GTX 1050 Ti has the same layout for core and memory as the higher end cards, because it would be nice to be able to pick up an Asus Strix block and waterblock a few of these lower end cards. Why lower end ones instead of the higher end ones? I want two cards in a rig, but can only use up so much power, and want to put a pair of them in alongside with a dual E5-2670 rig ^_^
I think the review embargo
)
I think the review embargo released some hours ago, and there are quite a few reviews posted. As most of them are quite firmly targeted at game-players, squeezing out information relevant to running Einstein is a bit iffy.
Some general points:
1. quite low power (at least one reviewer specifically called out slightly lower power than a GTX 750 at game settings)
2. extremely low power at idle (good if you don't BOINC 24/7, but I do, so this matters little to me).
3. Some greatly prefer to the 1050Ti to the base 1050, others say that, especially because clock rates come set faster on the 1050, that the extra performance of the 1050Ti is not likely worth the extra price and thus greatly prefer the 1050.
4. That last argument may crucially depend on how 4 vs. 2 Gbytes of VRAM actually affects you.
5. a dual-fan version for which I saw a review is really, really quiet.
6. a lot of the models of the cards are not widely available at all yet. In a quick look I did not spot any on offer at Amazon, and Nowinstock had not yet started a tracker.
7. for those of us who overclock, it is disconcerting that some hard upper limits on both memory and core clock speed were observed. As Einstein work is not likely to exceed the power limits, these artificial rate limits may stop us from getting more out of the cards. On the plus side, maybe fewer of our participants will run rigs returning high error fractions because of excess zeal and inadequate checking and monitoring.
I'm currently inclined to get one, but am not at all settled on which one. I'm pretty fan noise allergic (and have paid many extra pennies on more expensive case fans), so should probably get a dual-fan model. I currently have two 750Ti and one base-model 750 running, and would most likely do a direct swap, allowing comparison on the chosen platform.
I ordered an EVGA 1050Ti SC
)
I ordered an EVGA 1050Ti SC single fan model. All of my current 750Tis SCs are mostly single fan and run the same or cooler then my dual fan FTW 750Ti for similar clock speeds. I have one lil 750Ti that has an ASIC value of 89.9% and boosts almost 1400mhz on its own. Will the 1050Ti be the same? hope so since I like the simplicity of not needing the six pin connector. If I was going to OC the card further I would look into the dual fan ACX 3.0 version EVGA sells. I'm expecting a 20-30% increase in productivity with this card vs my exsisting 750s. I'm also a big fan that it comes with 7000mhz 4GB ram.
The whole announcement
)
The whole announcement availability structure seems a bit different for the 1050 than the earlier Pascal cards. As of now (a couple of days after release day) I still don't spot any listings on Amazon (not even the third party very high priced, buy now get shipped later stuff). NewEgg, on the other hand has several listings, and among the 1050Ti cards listed, many actually show as available for shipment. BH photo also has a few listings, but only for backorder.
My current intention is, mostly for curiosity, to buy a dual-fan 1050 (not Ti), prefering one that is short, cheap, and likely to have low and agreeable fan noise. I've backordered a Gigabyte GV-N1050OC-2GD, but that may well change if something I like better turns up before that ships.
By the way, I think people putting one of these as the only graphics card in a moderately well-ventilated case who are not unreasonably bothered by fan noise can safely get single-fan models. But for my usage none of these three conditions are met, so I choose to limit my choices or spend a little more to get two fans.
I have just ordered a 1050
)
I have just ordered a 1050 Ti. Was a bit expensive, but I couldn't resist. Should have it running on Sunday-Monday, will post results...
-----
Memory OCs help production
)
Memory OCs help production quite a bit with E@H. With the reduced mem speed and width I don't think these cards will scale down just based off the # of cores. I can easily push my 1070 to +900 MHz running E@H and every bit of it helps.