It maybe looks like a calculator, but it is a fully-featured ultra-mobile computer from the early 80s. Weighing just 315 grams, It has a custom 8-bit CPU, up to 16kB of RAM, built-in BASIC and display with a resolution of 160×32 pixels (20×4 characters). The machine can be expanded with some sort of a docking station that adds a four-color plotter and tape storage.
I took this photo while working on the MS QuickBasic version of our benchmark (previously written in C and Assembly). The goal was to compare speed of different languages and compilers. This required me to relearn BASIC so I always needed to see online help, my BASIC code and the original C code on a single screen. Switching a screen into the VGA 80×50-character mode is invaluable in these situations.
Regarding the results: I tested all versions of our benchmark, which is mostly about integer performance and memory access (similar to compiling and XML parsing), on this IBM PS/2 P70 with a 20-MHz 386 (DX). Interpreted QuickBasic 4.5 version was used as a baseline (1x). These are the speed-ups:
PowerBook 100 has a special place in my heart. I had one back in the 1990s and I loved its big trackball, comfortable keyboard and proper palmrest area – features that were not present in typical PC laptops. Other early PowerBooks were not as small and light as the PowerBook 100, but they shared many design decisions with it. I perfectly understand people who bought and used these computers when they were new. On the other side, I cannot agree with those who see early PowerBooks as universally superior machines to PC notebooks. That’s just not true.
From time to time, I see nonsense statements like that PowerBooks were the first laptops with stereo sound, optical drives, docking stations or other features. Not sure at the moment, but I think that you can find some of these statements even on Wikipedia. However, all of these features were previously available in PC laptops.
In fact, the generations of early PowerBooks that came after the first generation were not considered very innovative back then. Just a few examples:
The first color TFT PowerBook – 180c – was released in August, 1993 – almost a year after major PC brands released their first TFT portables. The PowerBook 180c was equipped with a small 8.4-inch 640×480 screen when PC laptops often used 9.5-inch screens and there were some with even 10.5-inch screens (like the famous IBM ThinkPad 700C – December, 1992). That was not the only issue – it lasted only about an hour on one charge because (unlike PC laptops) it didn’t have a 3.3V CPU, advanced power management features and NiMH batteries.
Heat and power consumption was so big issue with Motorola 68040 that Apple had to release 040-based PowerBooks with a version of the CPU that didn’t have a math coprocessor. Thus, programs that used it heavily were twice as fast when running on the previous generation of high-end 030-based PowerBooks. 486DX-based PC laptops could run the same code four times as fast.
The machines on the photos:
(1) Apple PowerBook 100 (1991), 16-MHz 68000, 4MB RAM, 20MB HDD
(2) Toshiba T2200SX (1991), 20-MHz 386SX, 4MB RAM, 60MB HDD
(3) Apple PowerBook 145 (1993), 25-MHz 68030, 8MB RAM, 80MB HDD
Our custom 3D-printed ultraSATAN installation kit for Atari STacy:
It might be surprising for somebody, but Itanium-based systems are still being used today. You cannot see them in a form of a standard office PC though. They are hidden in large datacenters, doing some serious stuff running on HP-UX or OpenVMS.
I have one HP Integrity BL860c i2 blade server (manufactured in 2010) in our lab and there was recently some time to play with it a bit. I played a lot with HP-UX before (even with PA-RISC machines) so the obvious next step was to install the last version of Windows Server for Itanium. It has a built-in x86 emulator written by guys from Intel and it is possible to transparently run standard 32bit x86 applications along with native ones. It is not very different from current ARM-based Windows devices.
Some simple emulated applications can achieve 50-70% of CPU performance. However, other applications (especially those with GUI) are super-slow. It takes ages to load a web page in emulated Firefox. The quad-core 1.33-GHz Itanium 9320 received only 0.38 pts in the Cinebench R11.5 multi-threaded test (32bit x86, 8 threads). That is similar to a single core of a five years old in-order Intel Atom processor.
It was expected to access Windows Server 2008 (R2) for Itanium no other way than using RDP. Therefore, there is literally no graphics acceleration implemented in the ATI driver (the chip is based on the ATI Radeon 7000 core).
The support for Itanium in Windows was discontinued in 2010.
SGI Indigo2 IMPACT systems were the best workstations for game development and other activities involving textured 3D rendering in 1995. My system is equipped with High IMPACT Graphics, which is a two-card solution with a dedicated geometry engine (one million triangles/s), raster engine with two pixel processors (two pixel per cycle, 60-70 textured Mpixels/s), 12MB of pixel memory and a single texture-mapping unit with its own 1MB of texture memory.
The high-end option was called Maximum IMPACT Graphics. It took three slots in the computer and doubled the rasterisation performance by using exactly the same principle that was later used by 3Dfx Voodoo2 SLI (scan line interleaving).
The 3D performance of SGI Indigo2 IMPACT was years ahead of PCs and other workstations. In fact, 3Dfx Voodoo2, the best gaming 3D accelerator for PCs in 1998, had similar performance to High IMPACT graphics but unlike the IMPACT series, it didn’t support windowed rendering, 32-bit color precision and high resolutions.
The last two photos show Indigo2 IMPACT systems during the development of Final Fantasy VII (source: Sony press kit).
GPUbench results – compare the SGI Maximum IMPACT performance with other 3D accelerators of the same era.
I was a big fan of All-in-Wonder cards back in the late-90s. ‘Multimedia’ was a big thing back then and ATI responded with the concept of a single card combining 2D/3D/video acceleration, TV-in/out and a TV tuner. This card allowed me to record and encode full PAL resolution in real time on a standard 266-MHz Pentium II PC. This one was based on the ATI Rage Pro chipset so it was not very good for 3D games. In following years, I had also All-in-Wonder cards based on Rage 128 Pro, Radeon and Radeon 8500DV. The last of them added also two Firewire 400 ports and a remote control. I loved it.
There is no demand for such things today. Almost nobody cares about the built-in TV input capability in PCs. In 2010, it was possible to buy a multimedia laptop with an integrated TV tuner and remote control. There were plenty of them (especially with 17.3 and 18.4-inch screens), but there is none today.
Although I’m not a collector, I always wanted to have one PowerBook 100 just to remind my childhood. It was the coolest Apple notebook from the early 90s. Last year, we brought back to life two out of three dead PowerBooks 100 from a friend of mine, so we could keep one for free. The restoration was not finished though. All the 2.5-inch Conner SCSI hard drives were dead.
These drives are hard to find so I installed the PowerMonster II adapter last weekend. I would do it maybe half a year ago but the first experiment was with a 2GB CF card and resulted in the “SCSI Bus Not Terminated” error. I was later told that old SCSI Mac don’t like large hard drives. With a 512MB CF card, the system immediately detected a new drive on the bus.
Making it usable with Mac was a different story. OS installer diskettes contain an utility for hard drive setup but it refuses to work with non-Apple drives. I tried even “hacked” versions that should work with other drives but with no success. Finally, after trying multiple utilities, Lido 7 rescued me five minutes before I wanted to give it up. I clicked on “Easy Setup”, set a name and icon for the drive and everything was ready to install the operating system.
Unlike DOS and UNIX systems, old Macs never had universal utilities to set up any third-party hardware. I don’t like that approach but Apple had that as a part of its strategy.
The game runs smoothly on the 16-MHz Motorola 68000 and has better music and sound in comparison with the PC version. Unlike other passive-matrix displays of the era, this 640×400 1-bit panel from Sharp is really fast and makes the game quite enjoyable.
This machine represented the SGI’s low-end workstation offering. It was targeted towards Mac users (DTP…) that needed more graphics and processing power than they could get from Macintosh Quadra systems. It’s a sleek pizza-box computer with just a single quiet fan inside (unlike other SGI systems). However, in comparison with non-SGI competitors, it was not slow. It had at least 100-MHz (64bit) MIPS processor, at least 16MB of RAM (reasonable configurations started with 32MB) and multiple graphics card options available. 10Mbit/s LAN, ISDN modem and video inputs (composite / S-Video / a digital port for the bundled webcam) were integrated on the logic board in all configurations as standard.
My Indy is from 1995 and has a more powerful 150MHz MIPS R5000 CPU. On the other side, it is equipped with the lowest possible graphics card (XL8/XGE8/Newport) that supports no more than 256 colors and was introduced with the early machines.
I always thought that Indy was the only SGI system without any 3D acceleration when sold with XL8 (2MB of 64bit video RAM) or XL24 (6MB of 192bit video RAM for true color modes) graphics cards. I expected just a crappy framebuffer (with BitBlt) and nothing more – like in Sun and HP machines. I was wrong. The REX3 chip inside the Newport graphics is pretty capable. Although all the 3D transformations and triangle setup are done in software, the chip can raster triangles with smooth (Gouraud) shading and per-vertex alpha-blending. Even Z-Buffering is partially accelerated using the chip (Z-Buffer is stored in system memory though).
In fact, this chip is not very far from early PC 3D accelerators (1996-1997) in terms of functions… except for the texturing support which was not available even with higher-end workstation-class 3D accelerators in 1993. This is for the first time I see 3D accelerated OpenGL (1.0) on such an old graphics card – and in 256 colors. To be correct, the scene with triangles has just 16 colors because any real-time graphics requires double-buffering. Each byte of the window in video memory contains one pixel from both buffers in the GBRG-GBRG arrangement of bits.
The graphics card is faster than I would expect. The pixel fill rate for smooth shaded triangles is ~50Mpix/s. If you add alpha-blending, you will get ~20Mpix/s. That’s 5-20 times as fast as the Windows NT 4.0 software renderer on a laptop with 133-MHz Pentium MMX and a 2D-only graphics chip. The speed in 3D is more comparable with 3Dlabs Permedia, S3 Virge DX and other consumer 3D accelerators from 1996.