It's expensive because the via stuff was mainly intended for industrial uses like kiosks. It's not really targeted at cheapo at home people. I had a Via C7 and a Via Eden both used for my firewall (dsl modem speeds at the time so the fact those are weak processors didn't matter). I used them because they produced so little heat that cooling in the metal box wasn't a problem (they had passive heat sinks)
Targeted only because the price / performance did not make sense for sane people. All embedded projects we had Via processors failed at a high rate. This was definetly not high quality stuff.
I'm watching this video and in 2019, there was already the Via Nano X4 series available. I got an Eden X4 C4450 in 2016. It seems the support for the VIA Nano X4 was in Linux till about 2020. O could boot in Ubuntu 16.4 and 18.4 but after that, it wouldn't allow me to install Ubuntu Conventionally. It's also not well supported by NAS OSes FreeBSD is a no go. Others tend to freak out when they come across the platform. I did manage to get Open Media Vault 5.5.11 to install on it though eventually.
For anyone who finds this today: that board has a PCI slot, not PCI Express. The Nano was the last chip series VIA put out before the Zhaoxin deal which isn't globally available. The chances of getting a better GPU than the Integrated one are almost zero.
Yeah, exactly my thoughts, guess one of the many reasons they were not that successful back in that day anymore I guess, together with the high asking price and not that stable graphics, ... PS: as I plugged that out in the vid for a test, I can not recumbent that for longer, as the heatsink became quite too hot to touch, so I rather suggest modding that for lower RPM instead, ..!
@@renerebe I think it's hard to compete in the x86 space. Even if you look at AMD, which is no small company ... they really had to give it their all to be competitive again. For a small company, unless you can find a niche, i think there is no chance ... In the bios, it said auto for the fans, but were there other options? Like quiet or low power or something?
detaart of course I checked the bios options, they are: Auto or Full speed! No kidding! ;-) via appeared to license that micro architecture to the Chinese Zhaoxin after that. PS: wish in general we had Linux fan register poking to tweak fan speed like that, also for any other board. Somehow that never was much of a thing I guess. Would have saved me a couple of reboots on server systems in the past, too, ..!
@@renerebe Haha, if this was auto, i'd hate to hear what "full speed" was like. Interwebs says 27.5w TDP ... so this is quite good for back in the day actually. It certainly does not need this fan, but then again ... too much for passive cooling too. The heatsink design seems weird too. Am i seeing this right? It seems like the socket is not inline with the fan, but rather offset to the side.
Hey Welcome! Luc of all people found my unsuccessful YT channel! ;-) Needless to say I could not get HDMI to output anything even after one hours of trying all sorts of fbdev or x drivers options and parameters, ... :-/
No miracle. Openchrome forked away from unichrome because i was apparently too rude in making sure that modesetting got anywhere (unichrome btw, is where the whole of structured display driver development, aka, modesetting, was pioneered). They wanted to have VBE as a secondary route for setting a mode. Having two paths ensures that users never report problems and that problems never get fixed. While i moved on to other things, like freeing ATI and freeing ARM GPUs, the openchrome people trundled along by sprinkling pci-ids over the codebase (enabling/disabling things if/when they worked, somewhat), and by even adding a third badly tested modesetting path. I am not sure whether they ended up doing a table deciding which modesetting path to use... Kevin Brace came in many years ago, and has been doing cosmetic changes ever since, acting as if he is rewriting the whole driver, but all of what he does is very superficial. His talk at XDC in 2017 is hilarious, mentioning my initials, but not mentioning me by name, and then going on to talk about exactly that what i pioneered and the openchrome people forked away for... I have 40+ devices here, but there really is no point in spending any time on this hw. VIA has been dying slowly ever since they introduced mini-itx back in 2003. Their revenue today is a fraction of what it was 16 years ago. The only reason it is not closed down yet is that it always has been a pet project of one of the Chen family, a wealthy .tw family that also owns or owned HTC. It's one of those things where it really is not worth anyones time. And it's not as if those VIA processor ever ran reliably. I have several dead devices where the assumption is that the CPU burned out, due to running too hot for too long. So do not expect this board to be used for anything much. If you really want a small efficient device, go to arm, there's OSHW and actual communities there.
Oh, and i have had to go claim copyright on my code twice there already. James Simmons, the guy who also poked at ps3fb, blatantly copy/pasted my code. And this was groundbreaking stuff like talking to the northbridge directly for sizing the reserved ram, and enabling a snooping path so that graphics memory accesses would not first have to pass through the actual gpu. I figured all that out during my coreboot bringup work (incidentally, VIA unichrome was the first hw to have a fully native display driver for coreboot, no binary blob required). I even enabled direct mapping on the AMD K8/Hammer, so that on amd-64 framebuffer reading/writing was lightning fast. Mr Brace mentioned not having fast access to the framebuffer on amd k8 during his 2017 XDC talk, and then Thomas Hellstrom (one of the guys who forked away from unichrome due to not accepting how important modesetting was) mentioned that "someone" worked around that, but he minced his words very carefully to not mention who or how.
@@lepetitdu17 I never did. Openchrome forked away from me. I had started xf86-video-unichrome when i noticed the turndown in xfree86, and i got the 2 other then active developers to come along. They later forked away because i did not accept VBE modesetting back in, as i predicted exactly that. They added a second path for native modesetting a few years down the line, making it 3 paths. And then introduced a system for disabling almost every feature (like hw cursor) until there were no more visible bugs (as then it would be all software). It took them until 2015 or 2016, when everyone had stopped caring for half a decade, until someone went round and threw that stuff away. Again, march/april 2005 is when it started, and i predicted it, yet no-one wanted to accept it. But then, most people in Xorg did not want to accept modesetting either, until they did, and then my name was no longer mentioned and my copyrights violated.
Rip Cyrix and Centaur. VIA helped Intel and AMD by buying those companies and not doing really anything with them......in the end all their low power x86 cpus got killed when everyone went to Arm or started to used Atom
@@renerebe I tried to buy a samsung via nano netbook back in 2009. It was sold out. I did a little research and find that a core 2 solo netbook is cheaper and faster. Also, the nano that i wanted has compatibility issues with everything other than windows. Linux or hackintosh is out of the question. no wonder it sold out quickly because no one wanted them.
It's expensive because the via stuff was mainly intended for industrial uses like kiosks. It's not really targeted at cheapo at home people. I had a Via C7 and a Via Eden both used for my firewall (dsl modem speeds at the time so the fact those are weak processors didn't matter). I used them because they produced so little heat that cooling in the metal box wasn't a problem (they had passive heat sinks)
Targeted only because the price / performance did not make sense for sane people. All embedded projects we had Via processors failed at a high rate. This was definetly not high quality stuff.
Cool, I want a little thing like that for a micro home server. :]
I'm watching this video and in 2019, there was already the Via Nano X4 series available. I got an Eden X4 C4450 in 2016. It seems the support for the VIA Nano X4 was in Linux till about 2020. O could boot in Ubuntu 16.4 and 18.4 but after that, it wouldn't allow me to install Ubuntu Conventionally. It's also not well supported by NAS OSes FreeBSD is a no go. Others tend to freak out when they come across the platform. I did manage to get Open Media Vault 5.5.11 to install on it though eventually.
I got a X8 centaur hauls prototype though ruclips.net/user/live1PSmbjKsXuE?si=2dLPtAIGgSREy24f
Very cool! Do you intend on doing any benchmarks with a discrete video card?
Well, not really as crappy as their open source drivers are? I guess crysis on Windows, maybe?
For anyone who finds this today: that board has a PCI slot, not PCI Express. The Nano was the last chip series VIA put out before the Zhaoxin deal which isn't globally available. The chances of getting a better GPU than the Integrated one are almost zero.
thank you for doing this. Sadly... regardless of blame; the price is what killed the Nano which is tragic.
yep
That fan! What were they thinking?!
Yeah, exactly my thoughts, guess one of the many reasons they were not that successful back in that day anymore I guess, together with the high asking price and not that stable graphics, ... PS: as I plugged that out in the vid for a test, I can not recumbent that for longer, as the heatsink became quite too hot to touch, so I rather suggest modding that for lower RPM instead, ..!
@@renerebe I think it's hard to compete in the x86 space. Even if you look at AMD, which is no small company ... they really had to give it their all to be competitive again. For a small company, unless you can find a niche, i think there is no chance ...
In the bios, it said auto for the fans, but were there other options? Like quiet or low power or something?
detaart of course I checked the bios options, they are: Auto or Full speed! No kidding! ;-) via appeared to license that micro architecture to the Chinese Zhaoxin after that. PS: wish in general we had Linux fan register poking to tweak fan speed like that, also for any other board. Somehow that never was much of a thing I guess. Would have saved me a couple of reboots on server systems in the past, too, ..!
@@renerebe Haha, if this was auto, i'd hate to hear what "full speed" was like. Interwebs says 27.5w TDP ... so this is quite good for back in the day actually. It certainly does not need this fan, but then again ... too much for passive cooling too. The heatsink design seems weird too.
Am i seeing this right? It seems like the socket is not inline with the fan, but rather offset to the side.
Probably something similar to what nvidia was thinking when they did the FX 5800 ultra.
Openchrome not setting a mode properly... I think i predicted that... hrm... 14 years ago ;)
Hey Welcome! Luc of all people found my unsuccessful YT channel! ;-) Needless to say I could not get HDMI to output anything even after one hours of trying all sorts of fbdev or x drivers options and parameters, ... :-/
No miracle. Openchrome forked away from unichrome because i was apparently too rude in making sure that modesetting got anywhere (unichrome btw, is where the whole of structured display driver development, aka, modesetting, was pioneered). They wanted to have VBE as a secondary route for setting a mode. Having two paths ensures that users never report problems and that problems never get fixed. While i moved on to other things, like freeing ATI and freeing ARM GPUs, the openchrome people trundled along by sprinkling pci-ids over the codebase (enabling/disabling things if/when they worked, somewhat), and by even adding a third badly tested modesetting path. I am not sure whether they ended up doing a table deciding which modesetting path to use... Kevin Brace came in many years ago, and has been doing cosmetic changes ever since, acting as if he is rewriting the whole driver, but all of what he does is very superficial. His talk at XDC in 2017 is hilarious, mentioning my initials, but not mentioning me by name, and then going on to talk about exactly that what i pioneered and the openchrome people forked away for...
I have 40+ devices here, but there really is no point in spending any time on this hw. VIA has been dying slowly ever since they introduced mini-itx back in 2003. Their revenue today is a fraction of what it was 16 years ago. The only reason it is not closed down yet is that it always has been a pet project of one of the Chen family, a wealthy .tw family that also owns or owned HTC. It's one of those things where it really is not worth anyones time.
And it's not as if those VIA processor ever ran reliably. I have several dead devices where the assumption is that the CPU burned out, due to running too hot for too long. So do not expect this board to be used for anything much. If you really want a small efficient device, go to arm, there's OSHW and actual communities there.
Oh, and i have had to go claim copyright on my code twice there already. James Simmons, the guy who also poked at ps3fb, blatantly copy/pasted my code. And this was groundbreaking stuff like talking to the northbridge directly for sizing the reserved ram, and enabling a snooping path so that graphics memory accesses would not first have to pass through the actual gpu. I figured all that out during my coreboot bringup work (incidentally, VIA unichrome was the first hw to have a fully native display driver for coreboot, no binary blob required).
I even enabled direct mapping on the AMD K8/Hammer, so that on amd-64 framebuffer reading/writing was lightning fast. Mr Brace mentioned not having fast access to the framebuffer on amd k8 during his 2017 XDC talk, and then Thomas Hellstrom (one of the guys who forked away from unichrome due to not accepting how important modesetting was) mentioned that "someone" worked around that, but he minced his words very carefully to not mention who or how.
it's a shame that you no longer develop openchrome
@@lepetitdu17 I never did. Openchrome forked away from me. I had started xf86-video-unichrome when i noticed the turndown in xfree86, and i got the 2 other then active developers to come along. They later forked away because i did not accept VBE modesetting back in, as i predicted exactly that. They added a second path for native modesetting a few years down the line, making it 3 paths. And then introduced a system for disabling almost every feature (like hw cursor) until there were no more visible bugs (as then it would be all software). It took them until 2015 or 2016, when everyone had stopped caring for half a decade, until someone went round and threw that stuff away. Again, march/april 2005 is when it started, and i predicted it, yet no-one wanted to accept it. But then, most people in Xorg did not want to accept modesetting either, until they did, and then my name was no longer mentioned and my copyrights violated.
hows the ps3 stuff coming
mostly works for me, patches welcome 🦊
Rip Cyrix and Centaur. VIA helped Intel and AMD by buying those companies and not doing really anything with them......in the end all their low power x86 cpus got killed when everyone went to Arm or started to used Atom
All those companies got bought because they were already at the time not competitive. VIA was way behind AMD/Intel back in the mid 2000's.
They were competitive for what they were, price vs performance. And nobody buys not competitive companies ;-)
@@renerebe I tried to buy a samsung via nano netbook back in 2009. It was sold out. I did a little research and find that a core 2 solo netbook is cheaper and faster. Also, the nano that i wanted has compatibility issues with everything other than windows. Linux or hackintosh is out of the question. no wonder it sold out quickly because no one wanted them.