There's also orcad files for riva128 and gerber files for voodoo 1

And the NV2 register reference...!

Nice find! Would be cool when the unpacked contents are uploaded.

Ooo, that V1000 API and microcode source sounds particularly intriguing, given the supposed programmability of that hardware.

Wow, lots of stuff there. Thanks for sharing!

vtbl64 -i Vendor1of4.113 -o out1 :

Reading compressed segment 6442, 239875930, 0
Decompress frame error, target 29673, bit_pos 28695

vtbl64 -i VSS1of2.113 -o out2 crash after line :

Reading compressed segment 54485, 3292245031, 0 Reading compressed segment 54486, 0, 0 Reading compressed segment 54487, 3292298 […]
Show full quote

Reading compressed segment 54485, 3292245031, 0
Reading compressed segment 54486, 0, 0
Reading compressed segment 54487, 3292298550, 0
Reading compressed segment 54488, 3292325329, 0

vtbl.exe Vendor1of4.113 :

found compression terminator in byte stream at 0xb671617
read 0x8 input bytes, generated 0x8 output bytes
fatal error: cum_sz 0xe4cafc7 != tot_wr 0xe4c458f

vtbl.exe VSS1of2.113 :

found compression terminator in byte stream at 0x3ea0ebf3 read 0x2d3f input bytes, generated 0x91d0 output bytes error obtaining […]
Show full quote

found compression terminator in byte stream at 0x3ea0ebf3
read 0x2d3f input bytes, generated 0x91d0 output bytes
error obtaining current output file position
during decompression 14048 segments required multiple calls to decomp_seg()

Did anyone successfully extract uncompressed images?

You can use the -s option, 11 segments are corrupted. We have most of the files extracted but we are working on basically producing a downloadable thing to put on IA.

starfrost wrote on 2026-09-04, 10:03:

You can use the -s option, 11 segments are corrupted. We have most of the files extracted but we are working on basically producing a downloadable thing to put on IA.

Original dtiom98.exe from IOmega Backup 98 can extract damaged archive ?

Edit:
Checked myself - it request last disk of backup set, so dtiom98.exe is pointless if you have only first part of backup set

possibly?? i might just put everything we found so far on gdrive

Played with IOMega's original dtiom98.exe:
1) VSS backup missed r-s-t-u-v-w-x-y-z folders, so SourceSafe probably missed a lot of files/history
Also many damaged in n\*, o\* and p\* folders

2) Vendor Back is OK, has few corrupted files:

ENGINEER-2\Engineer\Vendor Docs\Panda Project\VSPA12.PDF ENGINEER-2\Engineer\Vendor Docs\muRata\ferrite.pdf ENGINEER-2\Engineer\ […]
Show full quote

ENGINEER-2\Engineer\Vendor Docs\Panda Project\VSPA12.PDF
ENGINEER-2\Engineer\Vendor Docs\muRata\ferrite.pdf
ENGINEER-2\Engineer\Vendor Docs\Motorola\mc33201rev2.pdf
ENGINEER-2\Engineer\Vendor Docs\Mosel Vitelic\v53c16258.pdf
ENGINEER-2\Engineer\Vendor Docs\Molex\DVI\74320ad.pdf
ENGINEER-2\Engineer\VENDOR.SRC\nVidia\98apr\150e\Win96.zip
ENGINEER-2\Engineer\Vendor Docs\3Dfx\Voodoo3\new_mdk\B2_SG_AGP_MDK_REVC.zip
ENGINEER-2\Engineer\VENDOR.SRC\nVidia\98apr\146b\Resman.zip

we managed to recover a lot of stuff from the vss but basically what happened is that the guy only uploaded disk 1 of 2

98apr\146b\resman.zip is where 1 of 4 cut off

I had Claude make a standalone Python script that extracts all the contents and also gives a manifest + list files that are damaged.

Some very interesting stuff on here that has never been released before like the Speedy3D SDK! I also found a nVidia NV10 (Geforce 256) briefing to STB from October 1998 as well as alot of info on STBs MVP multi-monitor product with BIOS source and ROM images and info on the Permedia 3 among much more.

py iomega113.py --img "DISK.img" --out "OUTDIR" --jobs 12 --writers 1 --catalog Streams the extraction with a progress line, the […]
Show full quote

py iomega113.py --img "DISK.img" --out "OUTDIR" --jobs 12 --writers 1 --catalog
Streams the extraction with a progress line, then a JSON summary and any warnings, and writes two files next to OUTDIR:

<setname>.manifest.tsv — every entry: kind, full archive path, size, attributes, timestamp, and a note column flagging damaged or truncated files
<setname>.damage.json — the damaged-segment map, if any
Files land in OUTDIR\<image basename>\<drive>\

Look before you extract:
py iomega113.py --img "DISK.img" --catalog-only # the archive's own file list

--jobs N — decompression is a bit-level LZ77 loop and pure Python manages ~2 MB/s. Segments decode independently (I measured zero cross-segment history references), so this parallelises safely: --jobs 12 was 4.8× faster with byte-identical output. Use it always.
--writers N — Amount of files to write in parallell (as there are many many small files). Raise it for network targets, drop to 1 for a local SSD.

Ummm guys, are you sure that this is all that was uploaded from STB? Look at the user's uploads over the past few days:

https://archive.org/details/@shivashaw

cyclone3d wrote on 2026-09-04, 18:10:

Ummm guys, are you sure that this is all that was uploaded from STB? Look at the user's uploads over the past few days:

https://archive.org/details/@shivashaw

Are these safe to use? By safe I mean legal.

I have downloaded all the jaz images, zip archives etc right up until the... movie collection. From 10th of July 2026 until present . Had the clankers go through them. Besides the two images linked by starfrost in the first post, there doesn't appear to be anything of interest.

there was some stuff which was from various militaries (and we found the source code to the 1995 DOS game Stonekeep, and some vending machine code) but i don't think there is anything else from STB.

sharangad wrote on 2026-09-04, 18:17:

Are these safe to use? By safe I mean legal.

No

This is a random uploader uploading disks en masse. Nothing here is legally cleared and it'd be foolish for emulators to utilize them... It'd be the same "this will help emulators" naive bullshit i've seen whenever nintendo had a big leak. It's poison, even to glance it.

leileilol wrote on 2026-09-04, 23:27:

sharangad wrote on 2026-09-04, 18:17:

Are these safe to use? By safe I mean legal.

No

This is a random uploader uploading disks en masse. Nothing here is legally cleared and it'd be foolish for emulators to utilize them... It'd be the same "this will help emulators" naive bullshit i've seen whenever nintendo had a big leak. It's poison, even to glance it.

Unless it's something that nvidia take specific interest in and it could cause some conflict with their current commercial interests, it should be fine? That's I think why nintendo would have a problem with any historical leaks with their re-releases which often use their own emulators.
Most of these companies are defunct and the data is almost 30 years old

From what I can see there is no code or documents for any NVIDIA GPU after the RIVA TNT2, except for a single presentation (October 1998) about the GeForce 256, where they claimed it would be 800 percent faster