Apparently .... AMD has taken a Cue from Nvidia and they are artificially
"crippling" the double precision capability of the newer 290 290x series (((
That is why the 7970 - 280 is faster here on Einstein !
AMD was hoping we wouldn't mind forking over $4000 dollars for their 'Firepro'
GPU which has the exact same chip (Only not crippled )
Nvidia has been doing this for quite awhile also. An old 580 is faster than
a 680 for GPGPU and a 780 is just a crippled Titan chip which is in turn a
crippled Tesla chip.
Lots a cripplin goin on here (
I don't like it.
Bill
Quote:
Hello Bill,
I do have a single 290X running. I mainly run the card with FGRP Search currently which it performs well with.
With the other three Einstein projects, the 7970 is the better performer of the two cards although I am not sure why that is. For GW Search, the 290X can keep up with the 7970 running one task but with two tasks (0.5 factor), the 7970 pulls ahead.
Apparently .... AMD has taken a Cue from Nvidia and they are artificially
"crippling" the double precision capability of the newer 290 290x series (((
That is why the 7970 - 280 is faster here on Einstein !
AMD was hoping we wouldn't mind forking over $4000 dollars for their 'Firepro'
GPU which has the exact same chip (Only not crippled )
Nvidia has been doing this for quite awhile also. An old 580 is faster than
a 680 for GPGPU and a 780 is just a crippled Titan chip which is in turn a
crippled Tesla chip.
Lots a cripplin goin on here (
I don't like it.
Mornin Mikey,
I have an older (2011 model) kindlefire tablet a friend gave me and,
I use it almost every day. It is far more useful than I thought it
would be ! Best Buy says that tablet sales are now taking a dive over
the last couple of months. Everyone that wanted one has one (or 2 or 3)
tablets by now and, they aren't feeling a big need to upgrade.
Bill
Edit: Unlike us data crunchers who sometimes feel the need to upgrade
our GPUs every couple of months )))
Quote:
That would make sense give their current costs and capabilities. You can get a quad core chip in a tablet now and it can do pretty much anything a desktop can do, within reason, and is more mobile then a laptop while doing it. Given the cloud storage the need for huge harddrives for file storage is no longer needed for the tablets either, keep it in the cloud and both your desktop and your tablet can access it as needed.
Hmmmmm .... Gary Roberts stated in another thread, that the Einstein Apps
do not need or use Double Precision calculations ? (That surprised me )
You would think that the Einstein Boffins would be hoping for Aaaaah Accuracy ? ))
JKJK if they don't need it they don't need it )))
There must be some other reason for the Sub-Par performance of the AMD 290.
Please forgive all the GPU Jabbering here in Milestones.
It's been a little over 3 years 10 months since I was bitten by the bug and joined the Einstein project with a humble Q6600 quad core machine with a NVidia 9600GT GPU and look what happened...
I have slipped past the half billion!!!!!! and looking forward to completing the next :)
It's been a little over 3 years 10 months since I was bitten by the bug and joined the Einstein project with a humble Q6600 quad core machine with a NVidia 9600GT GPU and look what happened...
I have slipped past the half billion!!!!!! and looking forward to completing the next :)
BTW: I was impressed by your Wifi Scanner ! I don't use Linux
but, I use InSSIDer once in a while.
Bill
Thanks
and
Thanks.
The scanner worked great and then an "uplift" to one of the called functions changed the naming convention and format of returned data. I never got around to fixing it. But I could ....
Mornin Jeroen +
)
Mornin Jeroen + Samson,
Apparently .... AMD has taken a Cue from Nvidia and they are artificially
"crippling" the double precision capability of the newer 290 290x series (((
That is why the 7970 - 280 is faster here on Einstein !
AMD was hoping we wouldn't mind forking over $4000 dollars for their 'Firepro'
GPU which has the exact same chip (Only not crippled )
Nvidia has been doing this for quite awhile also. An old 580 is faster than
a 680 for GPGPU and a 780 is just a crippled Titan chip which is in turn a
crippled Tesla chip.
Lots a cripplin goin on here (
I don't like it.
Bill
RE: Mornin Jeroen +
)
RE: Mornin Mikey, I have
)
Mornin Mikey + All, Hmmmmm
)
Mornin Mikey + All,
Hmmmmm .... Gary Roberts stated in another thread, that the Einstein Apps
do not need or use Double Precision calculations ? (That surprised me )
You would think that the Einstein Boffins would be hoping for Aaaaah Accuracy ? ))
JKJK if they don't need it they don't need it )))
There must be some other reason for the Sub-Par performance of the AMD 290.
Please forgive all the GPU Jabbering here in Milestones.
Bill
It's been a little over 3
)
It's been a little over 3 years 10 months since I was bitten by the bug and joined the Einstein project with a humble Q6600 quad core machine with a NVidia 9600GT GPU and look what happened...
I have slipped past the half billion!!!!!! and looking forward to completing the next :)
Gav.
RE: It's been a little over
)
That is OUTSTANDING !
Keep it up.
Bill
50+ million. Where has the
)
50+ million. Where has the time gone???
RE: 50+ million. Where has
)
That is EXCELLENT Rob !
BTW: I was impressed by your Wifi Scanner ! I don't use Linux
but, I use InSSIDer once in a while.
Bill
RE: RE: 50+ million.
)
Thanks
and
Thanks.
The scanner worked great and then an "uplift" to one of the called functions changed the naming convention and format of returned data. I never got around to fixing it. But I could ....
Congratulations Gavin and
)
Congratulations Gavin and robl!