#sBitx Using Raspberry OS ARM64-bit (Debian 12) #sBitx
|
On Fri, Jan 12, 2024 at 02:30 AM, Zach Davis - AB0E wrote:
While the 2GB Pi is used by many the 32Bit OS does limit the RAM considerably. While the 4GB Pi may show that most of that 4GB is usable in reality it is not.All true, but really applies to data-intense applications like databases (think Oracle, etc) much more than sbitx. Back in the 1990s I worked on an operating system for the DEC Alpha, one of the first 64-bit CPUs.? The saying back then is you don't need 64 bits, but you do need 33!? We set world records on database performance largely because we could keep >4GB in a single address space, something very important to Oracle.? I remember the time we put 4GB of memory into an Alpha for the first time and it crashed.? It was only then we realized that we did not sell disk drives bigger than 2GB!? We had to build a RAID set just to hold the crash dump! One of the guys on our floor was Jon 'Maddog' Hall, who gave Linus Torvalds his first 64-bit system, a DEC "Jensen" with Alpha in it.? IMO, it's a big reason why Linux was 64-bit ready even in the 1990s. Yet, honestly, the big win for 64 bit CPUs in the sbitx application will probably be (a) more instructions and registers in the 64-bit instruction set and (b) more efficiency when operating on data in 64 bit chunks rather than 32. However, I don't think the sbitx software is mature enough for the difference to show a lot.? Search this thread for comments about use of double precision when single would do, not the most optimal fftw interfaces, etc.? Note I say "I don't think".? I have not gathered detailed data on this subject, but given how easy it is to find issues in the software, I'm pretty sure my guess is justified. ? -- Regards, Dave, N1AI |
|
On Tue, Jan 9, 2024 at 11:44 PM, JerryP wrote:
I couldn't get dual HDMI monitors to work so I had to comment out the dtoverlay=vc4-kms-v3d line in config.txt. Otherwise, sbitx would not run because of the dreaded simple.c error from alsa sound. It probably has to do with HDMI audio being available on the monitor but I don't know enough about the alsa sound system to fix it.On Wed, Jan 10, 2024 at 08:13 AM, Rafael Diniz wrote: Use this, Jerry:I believe the problem is that the sbitx_sound.c code uses the generic device name 'hw:0,0' assuming that it is the sound card for the radio data input/output.? This means you can't let other sound devices that may appear earlier claim that name.? The video stack loads incredibly early so it wins.? The simple.c assertion happens because the sbitx code is trying to set up a mixer device on top of HDMI audio instead of our codec and that doesn't work.? The change Rafael gave tell the video stack to not enable HDMI audio so it doesn't claim that name.? I have some code that uses specific names so we don't have to disable HDMI audio in the future.? I'll do a PR at some point in the future when things quiet down.? If someone else fixes it first, that's fine by me. My /boot/config.txt also comments out?dtparam=audio=on which disables the headphone jack audio.? I forget if this is mandatory or if it is something I did when I was trying out various other things.? I believe that what Rafael says is correct, these issues aren't specific to Bookworm.? -- Regards, Dave, N1AI |
|
Dave, Yes, 'dtoverlay=vc4-kms-v3d,noaudio' solved the problem. It would be nice if the Raspberry Pi Org put these things on their documentation. I've added this as a issue to add this to their config.txt documentation. On Sun, Jan 21, 2024 at 11:54?PM Dave, N1AI <n1ai@...> wrote: On Tue, Jan 9, 2024 at 11:44 PM, JerryP wrote: -- Jerry, AC9NM 手中的鸟在灌木丛中值两个 --
Jerry Ponko, AC9NM |
|
On Mon, Jan 22, 2024 at 09:56 AM, JerryP wrote:
It would be nice if the Raspberry Pi Org put these things on their documentation.I agree.? It would also be nice if the sbitx software was more mature and robust.? The config.txt change is not a solution, it is a workaround.? It's not like wanting to attach a HDMI monitor with speakers to a sbitx is a crazy idea.? You can't do that now because the sbitx code will crash if you do.? The problem has been described on this forum for months if not years already.? If we had a useful bug tracker and a community of developers updating it and the software, maybe this would have been fixed months ago.? If it had, maybe other people's concerns would also get addressed sooner.? I am hopeful that we will see an increase in the pace of improvement of the sbitx software in the near future.? ?We shall see if that happens or not. ? -- Regards, Dave, N1AI |
Dave, N1AI
JerryP