Standard Def Steve wrote on 2021-01-05, 22:13:

ragefury32 wrote on 2021-01-02, 20:23:

Standard Def Steve wrote on 2021-01-02, 18:56:

Jeez, did the Powerbooks actually use 60x right up until the Intel switch? That's terribad if true. I used to own a 17" G4 and always thought that it felt a little sluggish in Leopard, despite having a 9700 Pro. The CPU being stuck on the 60x bus would definitely explain it.

IIRC on the desktop side, the Yikes/PCI G4 was the last machine to use 60x; all of the AGP machines used the much faster MPX bus.

Whoops - I confused the 60x with the MPx. The Aluminum PowerBooks all use the Intrepid/KeyLargo setup, so they are 133MHz FSB machines with DDR RAM. Even with the MPx they are not exactly keeping up with the 266MHz FSB on the Athlon Bartons, much less the 400/533/800 MHz on the Netburst/Pentium-M - at that stage it’s like trying to bring a Tualatin to 2005 - it’s got some good points but definitely showing its age.

Yeah, Pentium M was probably the chip that convinced Apple to finally ditch PPC. I mean, in 2005 you could pick up a cheap Dothan-equipped Inspiron that would run circles around a $2500 Powerbook G4 while using less power. It was definitely not a good look for Apple.

Adding some L3 cache to the Powerbooks would've helped alleviate some of the memory bandwidth deficit without sucking down too much extra juice. There were quite a few instances where the giant 2MB L3 in the PowerMacs really helped with CPU performance, so I'm still kinda surprised that Apple didn't go that route with the high end 15" and 17" Powerbooks. But hey - at least they didn't further starve the G4 of memory bandwidth with integrated video!

Well, yes and no. Apple started looking at Intel when Motorola cancelled the PowerPC 7500 chip, the mythical 64 bit non-IBM PowerPC G5 project that started in 1999 and died in 2001. It was supposed to be the next great chip that was to underpin Apple’s next big growth spurt, but it didn’t work out. Supposedly the chip was never fully synthesized and IBM had to cut the POWER4 server chip down to meet Apple’s needs for a G4 successor. IBM didn’t have a suitable northbridge for Apple to use in the PowerBooks, the chip was never really designed to sip power, and at the same TDP/energy envelope needed for a PowerBook it was only about as fast as the Apollo8 G4s, so there’s no roadmap for growth from the AIM alliance. Both IBM and Motorola were using PowerPC for embedded applications (IBM did well fabbing PowerPC “little bro” variants for Nintendo/Sony/Microsoft and selling POWER “big bro” servers to its business customers, Motorola (soon Freescale) were tossing G4 chips into high end network switchgear). No one was really interested to make high cadence silicon chips for a brash little Silicon Valley computing firm constantly asking for faster chips.

The Pentium M was this side project Intel had in its Haifa design center in Israel that was not really meant to be the future of the company. It was originally supposed to be the foundation of a cheap, highly integrated x86 chip (Timna) with RDRAM for high volume low price PCs - think Atom before Atom and with a P6 core. The difficulty the Israeli team dealt with were high latency on the RDRAM so they had to spend serious effort making the P6 core work with the architecture - that involved maximizing L2 cache usage. The Timna project was later cancelled because RDRAM prices never came down enough for the low cost PC to work.

The performance of the original run of Norwood P4ms made Intel afraid of losing the Mobile market to Transmeta and Athlon XP-ms (the Thoroughbred/Barton cores were excellent performers and rather efficient too), which is why they told the Israeli engineers to take the P6 core, shrink it down, give it the Netburst FSB, better power management, take whatever they learn with getting Timna going, and to make it work. That’s how Banias came to being with its 1MB of L2 cache, and how it gave the Pentium 4s a real run for its money.

Side note - more than a few of the engineers on the Pentium-M project team in Israel ended up going to PASemi - PASemi was originally a PowerPC licensee that was supposed to present Apple with a better PowerPC chip (PWRficient) to be fabricated by TI - they didn’t (it happened right at the eve of the Intel transition) - Apple ended up working with them on a few embedded projects (iPod, iPhone, etc) and eventually bought them out. They are now the core of Apple’s in-house silicon design firm.

kjliew wrote on 2020-03-08, 10:11:

Intel GMA series were lame ducks.They were always seriously lacking in real games performance compared to NVIDIA/ATI IGP chipsets at the same time. They might look good in benchmarks but when it came down to actual games compatibility, drivers support and maturity, they always looked bad compared to NVIDIA/ATI IGP chipsets. Some may argue that no one was playing games with IGP, so why would anyone care. Well, it depends, NVIDIA/ATI IGP were usually good at past 2 generations games, but Intel were plagued with tons of compatibility issues. Intel HD Graphics were so much better and that was when Intel was serious in catching up in the Graphics department. When the graphics vendors reduced to 2-horse race, and one of them was your main competitor while the other was extremely arrogant to team up as partner, Intel had no choice but to invest on its own. It is interesting to see what Intel has to offer later this year when they introduce their in-house dGPU to compete with AMD and NVIDIA.

i never had a problem with gma graphics even better using the modded driver as long as it was using a core2duo cpu

just in some netbooks i used sometimes the atom cpu was bottle necking the gma graphics in some games

I suppose this is an old discussion but once I stopped upgrading in the Athlon 64 3500 times and switched to notebooks/netbooks I found that the iGPUs were much better than I thought when and if keeping expectations low and mostly retrogaming oriented. The problems came with the SGX line of the later Atom SoC iGPU that I tested a lot with Win 8.1 and theorically only oriented to Win 7 (Starter edition, not without a reason). I read a lot about what theorically went wrong with it and still nowdays using a mini-itx config with the GMA500/3600 gpu line and everything went wrong with that. Mostly I'd say for driver reasons, it's strange that only a single o.s. had support and the others on beta or unsupported results, not to mention a incredible Linux situation. And even the o.s. correct Win drivers didn't exactly helped a lot update after update. In comparison even for retrogaming I had better memories with the Rage Mobility of some Dell notebook I had or the Atom N270 chipset iGPU that had a more stable situation. The GMA3600 instead feel like a smartphone GPU with high end theorical specs, integrated with force into a x64 config but had to work into x86 mode cause only 32bit drivers were officially existing and at the benchmarks feel like something's wrong on the speed side without understanding really why. The only real reason I find for that is that its power demand should be almost not existing for a notebook or mini-itx board, something like 3 to 5 watts and coming from a smartphone concept probably not really oriented to desktop o.s. or complex gaming. The only interesting thing left for its times was the H264 decoding that in Win really worked good.
But still I find interesting such problems to the point I still test it sometimes to understand this iGPU. Win 7 would be better anyway compared to Win 8.1 32bit but I only have that.

Hi there. Hopefully someone has some experiences in AGP display-cache or so called AIMM on intel I815e.
https://www.intel.com/content/dam/support/us/ … cs/29835102.pdf

Do I need a Bios-Update if I put an AIMM display cache module in my AGP slot?
Can any AIMM-module be used, or are they branded. e.g. a dell module will not work on a fujitsu machine.

Thx!

AIMM modules out there, are universal. Only on chipsets that supports AIMM.

Cheers,

Thx! I ordered a bunch in poland. My module from UK produces several different errors.
Icons have "pink" background, not all but some. Some games seem to be not effected (Turok1 D3D) but others GLquake (OGL) or Unreal (D3D) suffer from moving black lines all over the screen (some sort of lightning issue)... we will see if the AIMM module from UK was defective.

Performance boost is excellent and very! notecible!

I hoped that the "local" 4MB would fix the Amithlon start, but it does not. Still only "textmodes" available and so actually no amithlon on that machine. What a pity!

Since the module is used for depth buffer only it should not cause such a color corruption.

sure? but it does and it's only if it's in the slot. I'll have to check this again...thx!

...otherwise there must be a defective igp...that would be very bad....

It's 100% the AIMM module. Look at the colored background on the icons and the moving black lines in 3D games D3D and OGL.

Unreal

Looking at a newer i810e datasheet I noticed the wording on buffers in the display cache isn't unambiguous.
If it can contain more buffers perhaps we should make a test at different resolutions to see which stays inside. This corruption may be ideal for that.

Under Dos I can't see it anywhere. In Windows it's noticeable on some icons and it depends on the game. Turok1/2 seem to be fine in all resolutions.
In Quake2 and Kingpin it's everywhere, regardless what options/resolutions I choose. In unreal it's on some places but not often.

If I had to make a guess....I think one of the both chips is faulty.

Intel offered some "Bios-Files" for AIMM and for NO-AIMM, but these are not bios-updates you can only load them into memory. Don't know for what that should be useful.

Got the AIMM modules from poland...they are working great! so my first one from the UK was faulty!

How much of any difference between with AIMM and without?

All I had was one board that supported one and was not gaming on it so I didn't know.

Cheers,

very noticeable. I have to check. Should be about 25%.