@outlawal2

I think that you need to contact Laurant from AAA technology in Luxembourg.

Merry Xmas. Great to see what still can be done with our little system

@khayoz

Jtag between cpld and xmos works well.

In case of an external debugger, the cpld needs to hi-Z some jtag signals and does buffer TDI as well.

I am afraid that this was never properly tested

If possible, you could try to lower jtag clock speed.

@khayoz

Not exactly sure what you did. This dongle is a xtag debugger which you've connected to your xtag header on de X5000?

And then tried to do some development with xTimeComposer?

If I would not have owned a X5000, I would have bought the A1222 for sure. In the same way that if I would have been able to buy an Amiga 4000 as teenager, I would not have owned an Amiga 1200.

The A1222 is quite capable for everything OS4.

For me, EUR 1000 for the third fastest Amiga ever build sounds like a bargain. Especially when you consider the price of ancient big box Amigas. Even today.

If multicore would ever materialize, it will consolidate its third place even further.

But unfortunately we live in a different world where aliexpress and rpi seems to be the benchmark for anything Amiga that is not 68k based.

@AmigaSociety

Yes, it is a bit cryptic. The thickness of the line to the x4 slot already suggests multiple lanes. Compared to the thin line of the x1 slots.

I know for sure because I use the x4 slot for my Nvme driver and can confirm that indeed 4 lanes are connected. But you can also check out the connector pinouts. There's only a remark for the x16 slot that lanes 4-15 are not connected. There is no such remark for the x4 slot.

@AmigaSociety

If you want to know something technical about your X5000 then you could also check out the X5000 TRM yourself. If you don't have it then you can download it here: https://amigaone.files.wordpress.com/2 ... /trm_cyrus_1-1-1_aeon.pdf

But to answer your questions:

1. The X5000 x16 slot is directly connected to the P50x0 with 4 PCIe2.0 lanes. This means a bandwidth of 2000MB/s

2. The x4 slot is connected with 4 PCIe2.0 lanes to a PCIe switch. Which is in turn connected with 4 PCIe2.0 lanes to the P50x0. So also 2000MB/s. But a bit higher latency due to the switch.

"Just" 2000MB/s is no issue at all. Various tests on PC websites reveal that the difference in fps between PCIe2.0 x4 (2GB/s) and PCIe3.0 x16 (~16GB/s) is just a few %.

@msteed

Quote:

Filer is open source: http://openamiga.org/?function=viewproject&projectid=20

@AmigaSociety

Nice vid!

You should also check out your "Power" pref. This preference tool allows you to switch the card between high and low power mode. High Power means more performance.

My setting is saved to High power mode. But for some reason, it fails to apply High power mode after a reboot.
Kas1e came up with a fix by forcing the Power pref to run once after reboot with adding the following line to the the user-startup:


It looks like your power pref is set a low.
I can run the same voxel edges at 120-200 fps in the preview window (vsync disabled). And about 20pfs fullscreen

My personal favourite shader is multipass->Full screen radial blurr. This one runs at ~80fps in full screen mode.

@AmigaSociety

I am away from my X5000 but I can remember that somehow my setup was also not able to display the resolution that I want. Iirc this has to do with some limitation within the gfx library. For some weird reason, my 3440x1440 only shows up at 61Hz and not at 60 Hz. So maybe you play with the refresh rate for your 2.5k resolution as well.

@AmigaSociety

When I switch between monitor or between HD and RX card, I do not have to update the monitor file

Make sure that the monitor file uses DDC for the screenmodes. Then the RX driver will be able to calculate applicable clocks frequencies.

@trixie

Quote:

Amiwest 2:37:05

@AmigaSociety

Reasons could be that the RX 580 has more gfx memory and is about 2.5 times faster. If you don't care for those numbers then the RX 550 is just fine.

Of course there are not much games yet where you would notice the difference except for DOOM3 or Quake 4 maybe.

Infighting in the qemu subscene.

Now you have finished you initiation rituals.

@kas1e

I alway use -O3. Never experienced any issue with it. But sometimes you need to use volatile to prevent the compiler from over-optimizing volatile code.

@joerg

Quote:

Without diskcache of course. I only have SSDs.

But as final test I've also enabled diskcache in my kicklayout. Limit is the same. But at least now I can start the Spencer demo. So the truncated partition is usable.

Now this is really it. Sorry for those who are stuck with sfs2. But I am not going to spend more time on it.

@joerg

Quote:

Ok understood. I'll stop looking into it further because I cannot fix SFS2 itself.

@joerg

Quote:
Ah ok. I thought that you were saying that in the context of MediaToolbox. Anyways, a 1-2TB partition doesn't function properly on a >2TB drive. I can copy files and read back from it but the files are corrupted.

Quote:
That is not enrtirely correct. I do see filesystems sending the TD_GETGEOMETRY. Can't didn't watch closely if that is the case for SFS2 as well. But in any case, TD_GETGEOMETRY64 is not used except for mounter.library.

Quote:

Both the WB Disk info window and the WB drive root window top bar show the 3.72TB /xxxMB used correctly in case of NGF\01. In case of SFS2, the same windows shows 1.86TB. Of course I do not know if this is somehow an internal issue between SFS2 and WB which exist for SFS2 and not for NGF\01. But the partition doesn't function correctly anyway. In contrary to 1TB drives or NGF\01. This is at least a hint that somthing is wrong.

Quote:
From what I can tell, it doesn't. But these things happen when you don't have the tools at your disposal to check every corner of your software. I had to solve an issue with TD_GETGEOMETRY as well to make the >2TB drives get recognized correctly. Wouldn't have noticed the bug without having such a drive myself.

@Raziel

There's also Jxfs. I do not know the story behind it and why it is obsolete now according to this page: https://wiki.amigaos.net/wiki/JXFS

It still runs just fine on my sam440.

I think you need a pre final edition version because the final edition version is read only

@joerg

I've bought a 4TB NVMe drive which arrived today. Unfortunately, I've ran into the same problem as @Raziel. The disk capacity is shown correctly in MediaToolbox. I can define partitions of > 2TB. But they cannot be formatted. With SFS2, I get a disk is full. With NGFS, I get an uninitialised disk.

Now I was expecting the following sequence from Mediatoolbox:
- Mediatoolbox issues the read capacity (10) SCSI command -> opcode 25h
- driver returns logical block address as ffffffffh because the last LBA exceeds 32bit
- Mediatoolbox issues the read capacity (16) SCSI command -> opcode 9eh
- driver returns logical block address to the last LBA (64bit value)

But this does not happen.

Instead the following happens:
- Mediatoolbox issues the read capacity (10) SCSI command -> opcode 25h
- driver returns logical block address as ffffffffh because the last LBA exceeds 32bit
- Mediatoolbox issues the TD_GETGEOMETRY command
- driver returns struct DriveGeometry

And now things get tricky. Because how to I fill out the 32bit struct DriveGeometry for a >2TB drive?

I can break up the total amount of sectors into C*H*S. But what do I have to do with dg_TotalSectors. Set it to 0?

Anyways, When I do break up the amount of sectors into CHS, I get the correct value for my disk in Mediatoolbox.
But like Raziel, I get those red dots next to "Last usable cyl." and "Blocks per cyl." . When you click on those values with a red dot. The fields to the right give the range (minimum, maximum). When I enter those maximum values, my disk size is reduced to 1.86TB. When I create a partition in this situation, the maximum size is also 1.86TB of course.
Now the format succeeds and the disk is usable.

There's also a TD_GETGEOMETRY64 with a struct DriveGeometry64 defined for the mounter library. Struct DriveGeometry64 doesn't work with CHS but with a 64bit dg_TotalSectors. So I've added support for this command and it's indeed used by the mounter library. But not by MediaToolbox or any filesystem.

So the question is if I am doing something wrong. Or if MediaToolbox is limiting the maximum disk size to 2TB?

EDIT1:
Ok, forget about NGF\01. Apparently I did something wrong (maybe I've selected the old NGF\00)

So I've ignored the red dots in expert mode and defined a 3.72TB partition with SFS2. With a blocksize of 512 bytes, I cannot format and get the "Disk is full" format error.
When I increase the blocksize to 4096 bytes, I can format the drive but the result is only a 1.86TB partition. Further increasing the blocksize doesn't help.
With NGF/01, I can successfully format a 3.72TB Partition. No matter the blocksize.

So MediaToolbox does a lot of weird things with displaying values. It even shows red dots where it shouldn't. But ultimately it's SFS2 which limits the partition size to 1.86TB.

Hopefully we might see NGF\01 for other machines then the X5000 some day, otherwise those systems are limited in partition size.

EDIT2:
SFS2 partitions do not work reliable. Executing something will result in a GR. Not even if I split my 3.7TB into e.g. 3x 1.2TB. NGF\01 works fine in any case.
So it looks that SFS2 will not work with >2TB drives.