What a waste to melt down these chips...

I saw that auction and inquired if there was a MII-433GP in the lot, but I was told there wasn't. I suppose people are buying up the gold chipped CPUs in anticipation of gold prices going even higher.

There is supposedly a 16 KB cache enhanced version of the AMD DX4-100 as well.

feipoa wrote:

I saw that auction and inquired if there was a MII-433GP in the lot, but I was told there wasn't. I suppose people are buying up the gold chipped CPUs in anticipation of gold prices going even higher.

There is supposedly a 16 KB cache enhanced version of the AMD DX4-100 as well.

Wasn't there an AMD DX4-100, which essentially is a downclocked 5x86? Should have the same cache if this is so.

If it were a dummy X5, I'm not sure how AMD changed the multiplier recognition post-production.

feipoa wrote:

If it were a dummy X5, I'm not sure how AMD changed the multiplier recognition post-production.

Afaik the 5x86 uses multi's of 3x and 4x. Perhaps these very late AMD DX4's have the 4x multi?

Iirc I have a very late production date AMD DX4 laying around but haven't gotten around to do some actual testing on these parts yet due to time constraints and other projects.

You own one of these,
http://www.chipdb.org/img-amd-a80486dx4-100sv16b-247.htm

or these,
http://www.chipdb.org/img-amd-am486-dx4-100v16bgc-251.htm

I think these chips were pretty uncommon. If you own one, could you check for the 2x setting working as a 4x setting? If that is the case, then perhaps these chips didn't qualify at 133 MHz, or perhaps that AMD was targeting some niche embedded market with the 100 MHz labeling.

Tetrium wrote:

Wasn't there an AMD DX4-100, which essentially is a downclocked 5x86? Should have the same cache if this is so.

Isn't the Am5x86 essentially just a 486 core anyway?

feipoa wrote:

You own one of these, http://www.chipdb.org/img-amd-a80486dx4...6b-247.htm or these, http://www.chipdb.org/img-amd-am486-dx4...g […]
Show full quote

You own one of these,
http://www.chipdb.org/img-amd-a80486dx4...6b-247.htm
or these,
http://www.chipdb.org/img-amd-am486-dx4...gc-251.htm

Actually, the two I own are:
The 120MHz SV8B "Enhanced" version with 8kb of Writeback cache.
http://www.cpu-world.com/CPUs/80486/AMD-A8048 … X4-120SV8B.html

The 100MHz NV8T version with 8k Writethrough cache, which I think came after the SV8B because Intel sued AMD over copying their Writeback technology.
http://www.cpu-world.com/CPUs/80486/AMD-A8048 … X4-100NV8T.html

I'm not sure how the 16kb cache versions slot in (though I'd love to have one). As for the multiplyer, I know when I accidentially set the 2x/4x jumper on the board on the 120 system(thinking "hey! DX4!"), she succesfully Posted as running as a 160MHz Am5x86 chip. I quickly turned it off because I didn't want to damage it, but it might warrant some further investigation.

DonutKing wrote:

Tetrium wrote:

Wasn't there an AMD DX4-100, which essentially is a downclocked 5x86? Should have the same cache if this is so.

Isn't the Am5x86 essentially just a 486 core anyway?

Yeah, like my above story, I accidentally clocked my Enhanced Am486DX4-120 at 160MHz, and the post test stated it was an Am5x86. All the AMD 5x86's are, are fast clocked 486's with tweaked L1 cache. The Am486's that were really underclocked 5x86 chips were the 'Enhanced' SV8B versions.

Cyrix is the only one who actually made a real Pentium-class P5-clone for the Socket 2/3 systems, they were literally cut-down 6x86's modified for the 32bit 486 bus, but they weren't on the market long due to 6x86's success, and are now getting quite rare.

maddmaxstar wrote:

Cyrix is the only one who actually made a real Pentium-class P5-clone for the Socket 2/3 systems, they were literally cut-down 6x86's modified for the 32bit 486 bus, but they weren't on the market long due to 6x86's success, and are now getting quite rare.

You forgot about Intel's Pentium Overdrive, it also works in Socket 3

And afaik the 8kb AMD 486's versions were older then the 16k versions.

Tetrium wrote:

maddmaxstar wrote:

Cyrix is the only one who actually made a real Pentium-class P5-clone for the Socket 2/3 systems, they were literally cut-down 6x86's modified for the 32bit 486 bus, but they weren't on the market long due to 6x86's success, and are now getting quite rare.

You forgot about Intel's Pentium Overdrive, it also works in Socket 3

And afaik the 8kb AMD 486's versions were older then the 16k versions.

Yeah, I wasn't thinking of the POD's as they were so rare, and were actual Pentiums rather than a 486-derivative. Then again, same can be said about the Cyrix.

As for the 8/16k Am486s... I'm finding more about this, I think my DX-120 CPU is a bit of an oddball. It's labeled as an SV8B, which should have 8k L1 and only support maximum of a 3x multiplyer, however I have it working (and I have pictures) at 4x33, which shows in the BIOS as an Am5x86-P75 at 133. Plus both at normal 486DX4 and X5 multipliers, she's showing in Speedsys and PCConfig as having 16k L1. I'm currently researching this to see what's going on.

It was based on my chip that I made the assumption that the Enhanced Am486DX4's were underclocked 5x86's, but I think I'm wrong on that and the oddity here is I have a freak of nature CPU that shouldn't exist.

maddmaxstar wrote:

Yeah, I wasn't thinking of the POD's as they were so rare, and were actual Pentiums rather than a 486-derivative. Then again, sa […]
Show full quote

Tetrium wrote:

maddmaxstar wrote:

Cyrix is the only one who actually made a real Pentium-class P5-clone for the Socket 2/3 systems, they were literally cut-down 6x86's modified for the 32bit 486 bus, but they weren't on the market long due to 6x86's success, and are now getting quite rare.

You forgot about Intel's Pentium Overdrive, it also works in Socket 3

And afaik the 8kb AMD 486's versions were older then the 16k versions.

Yeah, I wasn't thinking of the POD's as they were so rare, and were actual Pentiums rather than a 486-derivative. Then again, same can be said about the Cyrix.

As for the 8/16k Am486s... I'm finding more about this, I think my DX-120 CPU is a bit of an oddball. It's labeled as an SV8B, which should have 8k L1 and only support maximum of a 3x multiplyer, however I have it working (and I have pictures) at 4x33, which shows in the BIOS as an Am5x86-P75 at 133. Plus both at normal 486DX4 and X5 multipliers, she's showing in Speedsys and PCConfig as having 16k L1. I'm currently researching this to see what's going on.

It was based on my chip that I made the assumption that the Enhanced Am486DX4's were underclocked 5x86's, but I think I'm wrong on that and the oddity here is I have a freak of nature CPU that shouldn't exist.

How did you get that chip? It could be a remarked chip

maddmaxstar wrote:

Yeah, I wasn't thinking of the POD's as they were so rare, and were actual Pentiums rather than a 486-derivative. Then again, sa […]
Show full quote

Tetrium wrote:

maddmaxstar wrote:

Cyrix is the only one who actually made a real Pentium-class P5-clone for the Socket 2/3 systems, they were literally cut-down 6x86's modified for the 32bit 486 bus, but they weren't on the market long due to 6x86's success, and are now getting quite rare.

You forgot about Intel's Pentium Overdrive, it also works in Socket 3

And afaik the 8kb AMD 486's versions were older then the 16k versions.

Yeah, I wasn't thinking of the POD's as they were so rare, and were actual Pentiums rather than a 486-derivative. Then again, same can be said about the Cyrix.

As for the 8/16k Am486s... I'm finding more about this, I think my DX-120 CPU is a bit of an oddball. It's labeled as an SV8B, which should have 8k L1 and only support maximum of a 3x multiplyer, however I have it working (and I have pictures) at 4x33, which shows in the BIOS as an Am5x86-P75 at 133. Plus both at normal 486DX4 and X5 multipliers, she's showing in Speedsys and PCConfig as having 16k L1. I'm currently researching this to see what's going on.

It was based on my chip that I made the assumption that the Enhanced Am486DX4's were underclocked 5x86's, but I think I'm wrong on that and the oddity here is I have a freak of nature CPU that shouldn't exist.

According to the AM486 entry on Wikipedia

http://en.wikipedia.org/wiki/Am486

Later AMD 486 chips did get 16k of cache.

"The enhanced Am486 series supported new features like extended power-saving modes and a 8 KiB Write-Back L1-Cache, later versions even got an upgrade to 16 KiB Write-Back L1-Cache."

Tetrium wrote:

How did you get that chip? It could be a remarked chip

I scavenged the chip out of a pile of computer parts for Recycling at work a couple of years ago. I literally just saw it in a stack of chips and grabbed it for my 486. I don't know what it came out of before that.

sliderider wrote:

According to the AM486 entry on Wikipedia […]
Show full quote

According to the AM486 entry on Wikipedia

http://en.wikipedia.org/wiki/Am486

Later AMD 486 chips did get 16k of cache.

"The enhanced Am486 series supported new features like extended power-saving modes and a 8 KiB Write-Back L1-Cache, later versions even got an upgrade to 16 KiB Write-Back L1-Cache."

Yeah, but the 8k or 16k was denoted in the CPU's Product ID:
Am486DX4-120SV8B - Enhanced 486 with 8k Cache
http://www.cpu-world.com/CPUs/80486/AMD-A8048 … X4-120SV8B.html
Am486DX4-100V16BGC - Enhanced 486 with 16k Cache
http://www.cpu-world.com/CPUs/80486/AMD-Am486 … -100V16BGC.html

Mines clearly marked as a SV8B yet I've tried it in 2 different 486's, and it shows in both as having 16k cache, as well as if I try to clock it to 2x33 for DX2-66 speeds, the 4x multiplyer of the X5 kicks in and the chip is identified as an Am5x86-P75 at 133MHz. Very strange.

Here's photos:
Closeup of the chip:

And the PCConfig and Speedsys screens.

The production date is the 9th week of 1996, which is very late I think for a 486. Maybe after a certain date they all came with 16k cache regardless of what is stamped on the top of the chip or maybe AMD just remarked some 5x86 chips to fill demand for DX4's. Again, according to wiki, there should be some DX4-100 chips out there with 16k cache, too.

I went for a drive to my Parents place today, I picked up a couple more of my Retro PCs from my collection.

My AT&T PC-6300:

I don't know too much on the specs of it off the top of my head. It uses a proprietary Keyboard and Monochrome Monitor, as well as an original Microsoft Bus mouse to go with it. It has an 8086 processor and a hard drive, runs DOS, etc... it's an IBM PC Compatible, and not a full Clone, but it does run DOS, Windows 2, etc... As far as I know it works, but I haven't had it on in several years.

Here's a preview video I made on the AT&T and uploaded today: http://www.youtube.com/watch?v=mszM4JwFqx8

My IBM PS/2 55SX

Based on an 386sx-16, has an Intel 387sx-16 FPU, 16-bit MCA Bus, with a Logitech Bus mouse card and a Gravis Gameport card. (No sound! Waah!) Has a hard drive too. Not sure the exact specs on it since it's been in storage for almost 10yrs.

I also brought out my IBM PC/XT 5160, and my 286. I'll post more photos later.

Here's another one, one of my favorites but very old and grumpy and takes awhile to get working:

My IBM PC/XT 5160.

Built in January 1985, I bought this in September 1996 for 5 bucks at a corperate yard sale for my Dad's work. Today, it needs to be cleaned up a bit (scoured more like), the case has been slightly modified from it's original state as I used to have my 386sx "Spitoon" in this case. However, I kept everything and this machine is using all of it's original parts.

Specs:
8MHz AMD 8088 Processor
512k Total RAM
- 256k RAM on the motherboard.
- 256k RAM Expansion card installed via 8bit slot (full length card!)
Hercules Monochrome Graphics Card (Another full length card)
25pin Serial Port Card (with 25F-to-9M adapter)
360k 5.25" IBM Floppy Drive
31MB Seagate MFM Hard Drive (ST-238R I think)
Running MS-DOS 6.0

Unfortunately I don't have the original IBM Monitor, just a Packard Bell Monochrome monitor, very lightweight and no screen burn at least. I have a matching IBM Keyboard as well but one of the legs in the back is broken.

This was a bit of a bugger to get going. When I first powered it up (hasn't been on in over 4yrs) The RAM expanison card was flakey and the Hard Drive was spinning up and down inconsistently, returning the "1701" error code (She canna take the strain, Cap'n!) and boots to IBM Basic. Eventually I popped it open, reseated all the cards, and the machine booted normally with all RAM detected. So I installed a mouse driver and Monkey Island and played the slowest... game... ever.... it was like watching insects copulate.

maddmaxstar wrote:

Cyrix is the only one who actually made a real Pentium-class P5-clone for...

As far as online reading tells me, all Cyrix 5x86/6x86 designs were not Pentium clones; Cyrix used all their own designs and did not borrow from Intel licensing, nor did they reverse engineer the Pentiums. This is what excites me about Cyrix CPUs -- original thinking, the underdog, and clock-for-clock better than a Pentium. I think there was even a one month period of time where Cyrix had the fastest obtainable CPU. AMD K5's also spark my interest for this same reason.

feipoa wrote:

maddmaxstar wrote:

Cyrix is the only one who actually made a real Pentium-class P5-clone for...

As far as online reading tells me, all Cyrix 5x86/6x86 designs were not Pentium clones; Cyrix used all their own designs and did not borrow from Intel licensing, nor did they reverse engineer the Pentiums. This is what excites me about Cyrix CPUs -- original thinking, the underdog, and clock-for-clock better than a Pentium. I think there was even a one month period of time where Cyrix had the fastest obtainable CPU. AMD K5's also spark my interest for this same reason.

"Clone" might have been over-describing it, what I mean was that they're the same technological/performance level, like P5/6x86/K5 - Similar performing 5th Generation chips but all had very different architectures other than using Socket 5/7. Comparing that to all other socket 2/3 CPUs (other than the PODs) which were 486-generation chips. Cyrix's 5x86 was essentially a 6x86 modified for the 486 bus, whereas AMDs 5x86 as just a fast-clocked 486, hence why the performance was so badass on the Cyrix. I wish I knew that back in 2000 when I had my 5x86-100 chip 😜

I think you're right about Cyrix having the fastest CPU, I think that was mid-1996, 6x86-P166 or P200 or something was faster than Intel's chip in integer performance, and cheaper to boot. I liked Cyrix back in the day (I had a 6x86L-PR200+ and a Ti486DLC-40), a smaller player taking on Big Intel with a superior product, producing an outperforming processor at a lower clock speed. It's too bad they didn't keep up to date with their FPUs and weren't able to produce a good followup to 6x86. That was around the time I switched to AMD.

AMD's CPUs were custom design's from a certain point on as well. Their early stuff was cloned/reverse engineered, but everything AMD from K5 onward was custom. K5 could have been better if AMD were just able to get it to Clock faster, but it did provide some pretty decent performance-per-clock like the 6x86 did. Thankfully they bought NexGen with it's Nx686 chip which formed the basis for every AMD chip until Bulldozer. With that they've gone toe-to-toe with Intel for nearly 2 decades, K6, K6-III, Athlon and Athlon64 were all faster than anything Intel had upon initial release and they're still doing well thanks to buying up ATI (as much as I hate the fact they dropped the name, best integrated graphics on the market).

To get back on topic...

I posted a video of my PC/XT in action (bootup and running Monkey Island), and added it as a response to Lazy Game Reviews review of the PC 5150. He favorited it and retweeted it to his 24000 subscribers!

Here's the video link: http://www.youtube.com/watch?v=LofzYGATK9o