Login | Register For Free | Help
Search for: (Advanced)

Mailing List Archive: MythTV: Users

8400GS -> GT 430 -> ?? (60Hz, VDPAU High Quality)

 

 

First page Previous page 1 2 Next page Last page  View All MythTV users RSS feed   Index | Next | Previous | View Threaded


mythtv at theseekerr

Jun 20, 2012, 9:50 PM

Post #26 of 30 (843 views)
Permalink
Re: 8400GS -> GT 430 -> ?? (60Hz, VDPAU High Quality) [In reply to]

On 6/21/12, Gary Buhrmaster <gary.buhrmaster [at] gmail> wrote:
> On Wed, Jun 20, 2012 at 4:24 PM, Christopher Kerr <mythtv [at] theseekerr>
> wrote:
> ....
>> GT640's are just starting to become available as desktop cards - they're
>> based on chips intended for laptop usage, so I expect a low-profile
>> version
>> will be along shortly. They should be ~40% faster than GT440's (which are
>> basically just factory overclocked GT430's) which would make them strong
>> contenders as HTPC cards if a fanless version shows up.
>
> There is "faster" and there is "faster". For many applications, the
> GT430 is "faster" than a GT220, but not for running the de-interlacer
> code (the improvements are elsewhere). Any recent nVidia card
> can do decoding of the usual codecs (mpeg/h264) via dedicated
> hardware. The code that does the deinterlacing runs in the
> stream processors, and how many are needed, at what clock
> speed, is highly dependent on not only your content type but the
> nVidia stream processor architecture.

As you note, comparison is difficult. The GT430 has 96 Fermi-type
shader cores clocked around 1.4GHz. The GT640 has 384 Keplar-type
shader cores, which are substantially different and don't use an
independent clock (but for reference, they're clocked, along with the
core, at 900MHz).

Both have a 128-bit memory bus supporting DDR3 at 1.8GHz effective data rate.

So I'm optimistic, but right now I have no proof.

> I have seen no
> evaluation of the GT640M (the M is important) in regards to
> its capability

The M is less important for this particular card than you'd expect, as
the 640 uses the same (GK107) chip - the 640M just complicates matters
by throwing in a slower base clock with "Turbo" modes.

(To further complicate matters, there's also the GT630, which is an
OEM part which is basically the GT640M on a desktop card. Even more
confusingly, it used to be called the GT640, until nVidia renamed it
when they introduced the current GT640....)

> The nVidia triad for Linux HTPC: Low-profile, fan-less, able to
> perform Advanced 2X de-interlacing for all content in 60Hz
> countries. Pick any two (well, that is slightly unfair, since
> there is other criteria such as HDMI audio, and the presumption
> that fan-less does not include cooling solutions such as
> water heat transfer, and you are not going to do SMT rework
> to change the card layout, etc.)

I think this chip MIGHT be capable of hitting all of those points
provided you don't mind having a rather large passive cooler. Whether
anyone will build such a card remains to be seen.

- Chris
_______________________________________________
mythtv-users mailing list
mythtv-users [at] mythtv
http://www.mythtv.org/mailman/listinfo/mythtv-users


mythtv-list at dinkum

Jun 21, 2012, 12:52 AM

Post #27 of 30 (839 views)
Permalink
Re: 8400GS -> GT 430 -> ?? (60Hz, VDPAU High Quality) [In reply to]

On 21 Jun 2012, at 06:50, Christopher Kerr wrote:

> On 6/21/12, Gary Buhrmaster <gary.buhrmaster [at] gmail> wrote:
>> On Wed, Jun 20, 2012 at 4:24 PM, Christopher Kerr <mythtv [at] theseekerr>
>> wrote:
>> ....
>>> GT640's are just starting to become available as desktop cards - they're
>>> based on chips intended for laptop usage, so I expect a low-profile
>>> version
>>> will be along shortly. They should be ~40% faster than GT440's (which are
>>> basically just factory overclocked GT430's) which would make them strong
>>> contenders as HTPC cards if a fanless version shows up.
>>
>> There is "faster" and there is "faster". For many applications, the
>> GT430 is "faster" than a GT220, but not for running the de-interlacer
>> code (the improvements are elsewhere). Any recent nVidia card
>> can do decoding of the usual codecs (mpeg/h264) via dedicated
>> hardware. The code that does the deinterlacing runs in the
>> stream processors, and how many are needed, at what clock
>> speed, is highly dependent on not only your content type but the
>> nVidia stream processor architecture.
>
> As you note, comparison is difficult. The GT430 has 96 Fermi-type
> shader cores clocked around 1.4GHz. The GT640 has 384 Keplar-type
> shader cores, which are substantially different and don't use an
> independent clock (but for reference, they're clocked, along with the
> core, at 900MHz).
>
> Both have a 128-bit memory bus supporting DDR3 at 1.8GHz effective data rate.
>
> So I'm optimistic, but right now I have no proof.

I'm getting interested in this card too, there was an interesting (but not conclusive) review of one on Anandtech just now:
http://www.anandtech.com/show/5969/zotac-geforce-gt-640-review-

It's useful that they compare some benchmarks to the GT240 (640 beats it, just) which has shader performance massively beyond what 60hz countries require.

The reviewed card is single slot but with a noisy fan but I'd replace that with a big slow quiet top down fan, I'm sure there will be other cards along soon.

I think I may just get a GT240, I'm getting a 3D monitor soon and it will be native 120hz capable, a GT240 should be able to adv 2x de-interlace & IVTC 1080i60 into a 120hz frame buffer, if VDPAU makes this possible? Is that really advanced 4x or more likely advanced 2x x2 :-P


>
>> I have seen no
>> evaluation of the GT640M (the M is important) in regards to
>> its capability
>
> The M is less important for this particular card than you'd expect, as
> the 640 uses the same (GK107) chip - the 640M just complicates matters
> by throwing in a slower base clock with "Turbo" modes.
>
> (To further complicate matters, there's also the GT630, which is an
> OEM part which is basically the GT640M on a desktop card. Even more
> confusingly, it used to be called the GT640, until nVidia renamed it
> when they introduced the current GT640....)
>
>> The nVidia triad for Linux HTPC: Low-profile, fan-less, able to
>> perform Advanced 2X de-interlacing for all content in 60Hz
>> countries. Pick any two (well, that is slightly unfair, since
>> there is other criteria such as HDMI audio, and the presumption
>> that fan-less does not include cooling solutions such as
>> water heat transfer, and you are not going to do SMT rework
>> to change the card layout, etc.)
>
> I think this chip MIGHT be capable of hitting all of those points
> provided you don't mind having a rather large passive cooler. Whether
> anyone will build such a card remains to be seen.
>
> - Chris
> _______________________________________________
> mythtv-users mailing list
> mythtv-users [at] mythtv
> http://www.mythtv.org/mailman/listinfo/mythtv-users
>

_______________________________________________
mythtv-users mailing list
mythtv-users [at] mythtv
http://www.mythtv.org/mailman/listinfo/mythtv-users


jyavenard at gmail

Jun 21, 2012, 1:10 AM

Post #28 of 30 (823 views)
Permalink
Re: 8400GS -> GT 430 -> ?? (60Hz, VDPAU High Quality) [In reply to]

On 21 June 2012 17:52, Andre <mythtv-list [at] dinkum> wrote:
> I'm getting interested in this card too, there was an interesting (but not conclusive) review of one on Anandtech just now:
> http://www.anandtech.com/show/5969/zotac-geforce-gt-640-review-
>
> It's useful that they compare some benchmarks to the GT240 (640 beats it, just) which has shader performance massively beyond what 60hz countries require.

talk about overkill for an htpc...
_______________________________________________
mythtv-users mailing list
mythtv-users [at] mythtv
http://www.mythtv.org/mailman/listinfo/mythtv-users


Jens.Peder.Terjesen at devoteam

Jun 24, 2012, 11:48 PM

Post #29 of 30 (794 views)
Permalink
Re: 8400GS -> GT 430 -> ?? (60Hz, VDPAU High Quality) [In reply to]

-----Original Message-----
On 21. juni 2012 4:29 Gary Buhrmaster wrote:

On Thu, Jun 21, 2012 at 1:21 AM, Bob Long <bob [at] oblong> wrote:
....
> I wonder if there are any models that have the extra bulk of the
> heatsink on the rear side of the card? That typically may prevent a
> PCIEx1 card on the other side from being used, though.

I believe Sparkle had a couple of heat-pipe cards with the heatsink on the back a few years ago. I do not know of any company doing that today.

Gary
-----Original Message-----

Asus has a single slot GT 430 with heatpipe and heatsink.
http://www.asus.com/Graphics_Cards/NVIDIA_Series/ENGT430_DC_SLDI1GD3/

It is not low profile even though I have seen some websites claiming it is.
The strange thing is that the one I bought actually came with low profile brackets.

Jens

_______________________________________________
mythtv-users mailing list
mythtv-users [at] mythtv
http://www.mythtv.org/mailman/listinfo/mythtv-users


lists.md301 at gmail

Jul 16, 2012, 7:28 AM

Post #30 of 30 (685 views)
Permalink
Re: 8400GS -> GT 430 -> ?? (60Hz, VDPAU High Quality) [In reply to]

Following up on my previous reporting, for completeness and posterity's
sake:

On Mon, Jun 11, 2012 at 4:41 PM, lists.md301 <lists.md301 [at] gmail> wrote:

>
> I have a fanless GT430 myself (the above email and the capabilities helped
> make my decision), and for the most part used Advanced 2x with no problems
> (gentoo 0.24.1, nvidia-drivers-275.09.07, Samsung LCD-LED 55")...except:
> this has been my first baseball season with this configuration, and I've
> been noticing weird motion artifact during the pitcher's wind-up and the
> batters swing. I'm unsure if it's deinterlacing, or something else.
> (Never noticed any issues during football or hockey seasons.) Nothing
> obvious shows up in the verbose playback frontend logging. Just today I've
> upgraded to a more recent nvidia 295.49 driver, which I have yet to test
> (view playback). For what it's worth, I've also had audio sync issues with
> certain DVD rips with the Advanced 2x (Batman Begins, The Dark Knight),
> which were not present if I switched to Temporal 2x. I'll see tonight if
> the newer driver makes any difference.
>

I've been tweaking/struggling with this for about a month, but finally
resolved everything over the weekend:

(1) My upgrade to 295.49 did improve audio sync on ripped DVD's played
back with Advanced 2x deinterlacing, as well as a few odd picture anomolies
I've realized in hindsight (hockey puck motion weirdness). Audio sync is
still not absolutely perfect, but much better, that introducing an audio
sync delay isn't usually necessary (or the perceived delay offset doesn't
vary much, which is what I observed previously).
(2) The motion artifacts I observed previously were strictly a TV issue
(Samsung UN55D6050 LED-LEC). Took me a while to nail this down--last week
I finally hooked up my HDMI splitter, running the bigscreen and a 20" Acer
monitor (also capable of 1920x1080-60) and ran test observations of
problematic video simultaneously: baseball pitcher winding up, batter
taking practice swings. (I should also add the HDMI signal is first
passing through my Pioneer amp, which switches all my video sources, and is
connected to the Samsung on the HDMI/DVI input, but that input is not
configured for PC mode.) *** The artifacts were only on the Samsung.***
This tracked what I had previously observed on my Atom/ION2 systems
connected to smaller screens. So for the past week I fiddled with the
Samsung's picture modes. Choosing game mode solved it, but that seemed
like too big a hammer. Finally, I've found that setting "Auto Motion Plus"
(judder and blur reduction) to "Clear" eliminated the artifacts. "Clear"
is a distinct setting from "Off", and I don't know at what levels it puts
judder/blur controls. This net effect is similar to the global "game
mode", but without the loss/lock-out of other picture control functions.
(Another FYI, xorg is set up to choose the exact EIA modes for judder-free
playback, 59.94, 29.97, 24 fps etc, and the frontend playback is configured
to switch display modes using xrandr--per the wiki.)

Bottom line is, for my particular setup, I do not have any noticeable (to
me) video artifacts from my Zotac fanless GT430 video card as had earlier
feared. As always, YMMV.

First page Previous page 1 2 Next page Last page  View All MythTV users RSS feed   Index | Next | Previous | View Threaded
 
 


Interested in having your list archived? Contact Gossamer Threads
 
  Web Applications & Managed Hosting Powered by Gossamer Threads Inc.