Anyone remembered FX-4200?

[LabRat 891]

Their hands were probably tied, though, since making FX-4200 essentially would have admitted that Bulldozer was a failure. It's not CMT at all anymore. There were people who did the whole make-your-own-4M/4T thing with FX-8150 but pretty much all decided it wasn't actually worth doing even for gaming, once Vishera and and scheduler fixes came out.

Also, having zero viable higher end mini-ITX options for years...........those were some dark times.

By no means do I think FX's platform (or even that whole era) was "good".
IMO, It's just surprisingly not as bad as one would've expected, in retrospect.

If I end up w/ a FX-4200 and working-ish board, I'll try to (eventually) compare to the 3570k+z77 that I still have (also a 2500K, as parts).

 

[ShrimpBrime]

By no means do I think FX's platform (or even that whole era) was "good".
IMO, It's just surprisingly not as bad as one would've expected, in retrospect.

If I end up w/ a FX-4200 and working-ish board, I'll try to (eventually) compare to the 3570k+z77 that I still have (also a 2500K, as parts).

The FX lineup was directly competing for frequency.

The long pipelines where a copy paste of netburst.

So AMD had the first official 5ghz released processor. With the expense of IPC Having those long pipelines.

 

[fusseli]

Only slow on a general concensus.
They where much faster unzipping

So true! Well it can unzip files fastest in WinZip.... why again did I buy this? *shrug*

 

[ShrimpBrime]

So true! Well it can unzip files fastest in WinZip.... why again did I buy this? *shrug*

Lol. I had a lot of fun on that platform. Did a bit of dry ice and lN2. Was good for thrills!!

Not much of a gamer myself.

 

[fusseli]

By no means do I think FX's platform (or even that whole era) was "good".
IMO, It's just surprisingly not as bad as one would've expected, in retrospect.

If I end up w/ a FX-4200 and working-ish board, I'll try to (eventually) compare to the 3570k+z77 that I still have (also a 2500K, as parts).

It was bad. Like way more advertising over "bulldozer is coming!", worse performance, and off the heels of the Phenom disaster that somehow later turned out with the X4 Phenom 2s being faster than core2quads at the time.

Let's just all agree these are better times

 

[LabRat 891]

The FX lineup was directly competing for frequency.

The long pipelines where a copy paste of netburst.

So AMD had the first official 5ghz released processor. With the expense of IPC Having those long pipelines.

You've more/less confirmed the exact opinion I've had,
since I decided to "lolno, I'm out"@amd after my X3 720.

It was bad. Like way more advertising over "bulldozer is coming!", worse performance, and off the heels of the Phenom disaster that somehow later turned out with the X4 Phenom 2s being faster than core2quads at the time.

Let's just all agree these are better times
View attachment 325461

But it's only 4times faster than the AMD!

I remember the hype; 'watched a lot of promo materials on the Opterons based off it too.

IMO,
AMD deserved to be sued and lose; even if the case was nonsense at times (US courts, and all)

 

[Bones]

That particular HWbot run appears to be a core-reduced OC-run.
edit 2: it also appears to be a "single module, dual-core". IIRC, there was some configurability on the core-layout via firmware.
Agreed that it's 'odd'; I wouldn't have expected a 4xmodule 4xcore 1:1 module:core 'stock' CPU to 'unlock+lockdown' to a 1x module 2x core CPU...

• Model: AMD FX-4200 'Zambezi'
• Cooling: Single Stage Phase Change
• Cores: 5,002.49 MHz (+51.59%)
• View attachment 325454

It is.
Same as I did with mine while we were messing around at that time with these.

You can think of each FX physical core being like a coin - You have a front and obverse side to each one, these two sides representing a thread = 2 threads per physical core.
That's a typical Zambezi/Vishera setup in how their cores operate.

What makes the 4200 different is it does have 4 physical cores working but only 1 thread per physical core operating as if the "Coin" only had one image instead of the usual 2 you'd normally see.

What Splave and I did was to disable 2 cores in the BIOS (3 and 4), rebooted and ran the chips that way.

 

[ShrimpBrime]

You've more/less confirmed the exact opinion

IMO,
AMD deserved to be sued and lose; even if the case was nonsense at times (US courts, and all)

Welcome!!

The only thing about the AMD lawsuit that bothered be is that the FPU is considered a core, not the integer core it's self, which each cpu had 2/4/6 or 8 integer cores depending on the model.

FPU is just a core architecture defined separately from an integer core.

Essentially, the cpu has 12 cores in total (FX-8350 as example) 8 integer and 4 FPU. These process different instruction sets.

But because the normal was 8 integer and 8 FPU it was scrutinized and claimed false advertising. Which with 8 integer and 8 FPU cores, you have 16 cores total, not just 8... or in AMDs case using hyper transport to connect 4 modules which would contain "essentially" 3 cores each.

Since popular vote and eye witness wins most BS court cases, AMD lost because of the lack of understanding what the word "core" really means.

What was a cpu before the FPU was even invented? A bunch of transistors, nothing more.

An example of a cpu without floating point was before 80486. Quark D2000. But because it didn't have an FPU, it wasn't a cpu..... haha.

 

[natr0n]

I liked my phenom II more x4 be / x6 be

I did get 2 free fx 6300s a few years back.

felt sluggish and lackluster

 

[Sabotaged_Enigma]

Okay, I think we have the nail to the coffin.

Here's a snapshot in the BIOS of Asus M5A97.
It says "cache per core": L1 inst. cache: 64 KB, L1 data cache: 16 KB, L2 cache: 2 MB.
idk man anyone who has an M5A97 and another FX CPU to test and see what'll be displayed?
btw interestingly, FX-4200 is not even on the support list of this mainboard.

And in AIDA64, both version 6.60 and the latest 7.00, this is how it reports FX-4200: L1 code cache: 64 KB per module, L1 data cache: 16 KB per core, L2 cache: 2 MB per module.
But look at CPUID in both versions. There's no way that's correct. But I guess there's lack of "per xxx" at the end. They're the same when pure number is compared.
Unfortunately, AIDA64 doesn't report how many modules FX-4200's got. Otherwise it'll be another strong evidence.

So, I suppose my guess that FX-4200's got
320 KB L1 cache (256 KB L1 inst. and 64 KB L1 data) and 8 MB L2 cache
is confirmed? Can we agree with that?

@T4C Fantasy Hey bro, can we confirm this and maybe add it into our database? I can send you photo of my FX-4200 if you need it.

 

[Sabotaged_Enigma]

Did a bit of overclocking. Ended at 4.5 GHz. It could go higher, but I just didn't want to get it more than 1.45 V.
CPU-Z validation: AMD FX @ 4514.25 MHz - CPU-Z VALIDATOR. mT score over 1000, even greater than my Athlon X4 760K at 5 GHz...

Comparison time:
Cache and memory test, I pinned FX-4200 (blue stats), comparing to FX-8120 and FX-8350 at the same 4.5 GHz, you'll find FX-4200 has nearly been neck-and-neck.
GPGPU test, FX-4170, 4200, 6300, and 8350 together. Sorry these are all the models I have collected. I'd be grateful if you can give me some more lol.

 

[Sabotaged_Enigma]

There's one thing I'm still confused about. Why is FX-4200 even better than FX-8350 at FP32 and FP64, and why are they equivalent at INT performance? They've got the same number of FP units, and FX-8350's got double the number of integer units.
Scheduler drags efficiency down?

 

[ShrimpBrime]

There's one thing I'm still confused about. Why is FX-4200 even better than FX-8350 at FP32 and FP64, and why are they equivalent at INT performance? They've got the same number of FP units, and FX-8350's got double the number of integer units.
Scheduler drags efficiency down?
View attachment 331681

What operating system are you testing with? They where produced when W7 was most popular and W8 released in 2013. Vista was good after service pack 3, and I'm not sure about W10 or 11 because I ditched the platform long before the release of either, so no experience there.

 

[Sabotaged_Enigma]

What operating system are you testing with?

win10. Don't think it's caused by OS. More like a hardware-level difference.