Become a MacRumors Supporter for $50/year with no ads, ability to filter front page stories, and private forums.
The Amiga was used for video work and the Apple was on an Eye straining little screen, […]
Compared to the Amiga’s terribly flickering interlaced high-res modes, that little screen was probably a godsend. Small CRTs flicker less noticeably at a given refresh rate than bigger CRTs, and if the early Mac’s monochrome CRT uses long-persistence phosphors the flickering is even less noticeable.

Which components would be used?
CPU (The datasheet will contain loads of information)
What CPU would the accelerator board have?
 
Last edited:
  • Like
Reactions: B S Magnet
I reckon if you could use a power 5 or 6 CPU as these should be cheap to acquire, the board would be designed as an open-source type project and folk would source their own CPU.

Depending on what speed they achieve and also the bottleneck if any with the G5 produced if any.

The faster ram would increase performance but you would probably need some sort of cache for the data to return to the G5 through as not to bombard the PCIE with too much data at once.
 
Last edited:
  • Like
Reactions: pc297
I reckon if you could use a power 5 or 6 CPU as these should be cheap to acquire, the board would be designed as an open-source type project and folk would source their own CPU.

Depending on what speed they achieve and also the bottleneck if any with the G5 produced if any.

The faster ram would increase performance but you would probably need some sort of cache for the data to return to the G5 through as not to bombard the PCIE with too much data at once.
I am going to try some time to download the datasheets for the Power 5 and 6 Cpu's and look at the information to see if this is feasible!

I will try a draw up a basic diagram of how this might work. I just need to find the time to start my research.
 
  • Like
Reactions: pc297
ibm-power5-power9-roadmap.jpg


Here is the IBM roadmap

The POWER5+ is an improved iteration of the POWER5 introduced on 4 October 2005. Improvements initially were lower power consumption, due to the newer process it was fabricated in. The POWER5+ chip uses a 90 nm fabrication process. This resulted in the die size decrease from 389 mm2 to 243 mm2.

Clock frequency was not increased at launch and remained between at 1.5 to 1.9 GHz. On 14 February 2006, new versions raised the clock frequency to 2.2 GHz and then to 2.3 GHz on 25 July 2006.

The POWER5+ was packaged in the same packages as previous POWER5 microprocessors, but was also available in a quad-chip module (QCM) containing two POWER5+ dies and two L3 cache dies, one for each POWER5+ die. These QCM chips ran at a clock frequency of between 1.5 and 1.8 GHz.
 
Last edited:
  • Like
Reactions: dextructor
As of 2008, the range of POWER6 systems includes "Express" models (the 520, 550 and 560) and Enterprise models (the 570 and 595).[14] The various system models are designed to serve any sized business. For example, the 520 Express is marketed to small businesses while the Power 595 is marketed for large, multi-environment data centers. The main difference between the Express and Enterprise models is that the latter include Capacity Upgrade on Demand (CUoD) capabilities and hot-pluggable processor and memory "books".

IBM POWER6 (Power Systems) servers

NameNumber of socketsNumber of coresCPU clock frequency
520 Express244.2 GHz or 4.7 GHz
550 Express484.2 GHz or 5.0 GHz
560 Express8163.6 GHz
5708164.4 GHz or 5.0 GHz
57016324.2 GHz
57516324.7 GHz
59532644.2 GHz or 5.0 GHz
IBM also offers four POWER6 based blade servers.[15] Specifications are shown in the table below.

IBM POWER6 blade servers

NameNumber of coresCPU clock frequencyBlade slots required
BladeCenter JS1223.8 GHz1
BladeCenter JS2244.0 GHz1
BladeCenter JS2344.2 GHz1
BladeCenter JS4384.2 GHz2
All blades support AIX, i, and Linux. The BladeCenter S and H chassis is supported for blades running AIX, i, and Linux. The BladeCenter E, HT, and T chassis support blades running AIX and Linux but not i.

At the SuperComputing 2007 (SC07) conference in Reno a new water-cooled Power 575 was revealed. The 575 is composed of 2U "nodes" each with 32 POWER6 cores at 4.7 GHz with up to 256 GB of RAM. Up to 448 cores can be installed in a single frame.


IBM POWER6 disk storage
NameNumber of coresCPU clock frequencyNumber of controllers
DS87002, 44.7 GHz1, 2
DS88002, 4, 85.0 GHz1, 2
 
  • Like
Reactions: dextructor
If IBM had only got their act together and produced this instead of the 970mp!
 
Last edited:
I am now believing that the reason for the slower clock speeds is because IBM could not get rid of the heat in the CPU.
A massive piece of copper, for passive cooling must work to some extent but there does not appear to be any fans cooling the copper.

I wonder if the cost of manufacturing the PPC 970mp heatsink out of copper was too expensive?
 
Last edited:
@simie You're not the only one to look into just what made the POWER4, 5, and 6 tick, and as you can see, IBM wasn't particularly interested in the PowerPC 720. POWER4 just wasn't meant to go that fast, and look at what Apple had to do to get their variant based on the same tech go that fast. POWER5 seems to have really been more of a general upgrade and not an overhaul, except for one major issue, it uses an internal memory controller. Well that, and the fact that P5+ added ddr2 support.

Even if the G5 board could support these chips somehow, I really do not think the tech lines up, and that's why we never saw any drop-in cpu upgrades from the likes of Sonnet, who had been making a business out of it 1986, if you believe Google.
 
@simie You're not the only one to look into just what made the POWER4, 5, and 6 tick, and as you can see, IBM wasn't particularly interested in the PowerPC 720. POWER4 just wasn't meant to go that fast, and look at what Apple had to do to get their variant based on the same tech go that fast. POWER5 seems to have really been more of a general upgrade and not an overhaul, except for one major issue, it uses an internal memory controller. Well that, and the fact that P5+ added ddr2 support.

Even if the G5 board could support these chips somehow, I really do not think the tech lines up, and that's why we never saw any drop-in cpu upgrades from the likes of Sonnet, who had been making a business out of it 1986, if you believe Google.
The thing is for us to get a Power 5 or 6 server to run OSX on, they are not cheap!
Apple has prevented the development of X86 OSX which limits hackintosh folk by their M Series CPU's but at the same time they have hiked up the prices.
 
Last edited:
The thing is for us to get a Power 5 or 6 server to run OSX on, they are not cheap!
Apple has prevented the development of X86 OSX which limits hackintosh folk by their M Series CPU's but at the same time they have hiked up the prices.
See, that's the thing. You're asking about a solved problem. Oh, sure, it's using QEMU, but it's also using KVM, which means it's a virtual machine, not unlike running ARM Windows on an M1 Mac. There's extra steps because the POWER9 is a little different from the old PPC chips of Apple's old hardware, but if you got your hands on a POWER5 or 6 system, I would imagine the steps involved would be pretty similar.

Yes, you can use QEMU to run just about any architecture anywhere you can compile it, but this is mostly not emulation.
 
I have been looking around IBM's website and I have come across some information that I want to share regarding hardware acceleration.

I will post the information it on here shortly.
 
Not what I promised but an interesting read


 
Last edited:
Anyone tried with registered memory yet? Those 8Gb -R or -P sticks will give the telephone beep when used alone, but I will try to dump DIMM info with those in combination with 1-2Gb DIMMs in the first two slots. In theory the G5 Northbridge supports registered memory

Also, do the first two DIMMs have to be UDIMMs or can they be ECC? Will give it a try with either.

Re the OF script, can it be placed in ofboot.b?

Cheers,
 
I am going to try some time to download the datasheets for the Power 5 and 6 Cpu's and look at the information to see if this is feasible!

I will try a draw up a basic diagram of how this might work. I just need to find the time to start my research.
I would have a couple of Power7+ CPUs to donate for this project if it's of any help
 
View attachment 2097776
Core 2 Duo is socket 775 and the Core Duo microprocessors use Socket BGA479 and Socket M.

Some laptops are better quality than desktop machines, we have to look at the components inside like the ram (the speed and the quality of it, the quality and read and write speed of the HDD and how the device dissipates the heat.

I am glad you reminded me of the cpu types.

You said it there is no point in comparing processors of different types Intel is not a PowerPC cpu and is design as a market filler, Intel is an enormous company who are after profit. Apple is after profit, IBM made a processor that is very capable but could not keep up with Apples demands. Did Jobs ever want Apple to go in the direction that it is heading?

The early Intel Xeons MacPro's could not match the PowerPc 970mp, I proved this years ago with the Photoshop Test.

There will always be a bottleneck with any design in order to cut costs.
Socket M for sure, but were there any PGA479 Yonah's as in the same as for Pentium Ms? Intel's nomenclature is very misleading...
 
Has anyone managed to get a compression accelerator to work in a PPC - if you could get one to work then the performance within OSX should improve and take the strain off of the PPC970.

Ref IBM https://www.ibm.com/docs/en/linux-o...i-pcie3-fpga-compression-accelerator-adapters
Compression Accelerator Adapter

The accelerated GZIP cards implement the well-defined, open standard DEFLATE compressed data format, which is used in zlib, gzip, Java, and many other applications. Within the gzip and zip file formats, it has become the standard for compressed data exchange.


Does the software compression version exsit within OSX like in Linux and can this speed be improved.
Some folk have Linux on the PPC's so why not look and see if this could be improved, see the link above.
I currently do not have linux installed.
 
Last edited:
  • Like
Reactions: sasho648
Anyone tried with registered memory yet? Those 8Gb -R or -P sticks will give the telephone beep when used alone, but I will try to dump DIMM info with those in combination with 1-2Gb DIMMs in the first two slots. In theory the G5 Northbridge supports registered memory

Also, do the first two DIMMs have to be UDIMMs or can they be ECC? Will give it a try with either.

Re the OF script, can it be placed in ofboot.b?

Cheers,
I will try over the weekend with some registered dimms as I have some on my desk doing nothing.
 
  • Like
Reactions: pc297 and sasho648
Anyone tried with registered memory yet? Those 8Gb -R or -P sticks will give the telephone beep when used alone, but I will try to dump DIMM info with those in combination with 1-2Gb DIMMs in the first two slots. In theory the G5 Northbridge supports registered memory

Also, do the first two DIMMs have to be UDIMMs or can they be ECC? Will give it a try with either.

Re the OF script, can it be placed in ofboot.b?

Cheers,
8GB sticks would require more work FYUI - if even possible. As I've said the limit is 4GBs and even that takes an 0x10000000 bytes hit.
 
Has anyone managed to get a compression accelerator to work in a PPC - if you could get one to work then the performance within OSX should improve and take the strain off of the PPC970.

Ref IBM https://www.ibm.com/docs/en/linux-o...i-pcie3-fpga-compression-accelerator-adapters
Compression Accelerator Adapter

The accelerated GZIP cards implement the well-defined, open standard DEFLATE compressed data format, which is used in zlib, gzip, Java, and many other applications. Within the gzip and zip file formats, it has become the standard for compressed data exchange.


Does the software compression version exsit within OSX like in Linux and can this speed be improved.
Some folk have Linux on the PPC's so why not look and see if this could be improved, see the link above.
I currently do not have linux installed.
That might work into Linux - do tell if it does.
 
8GB sticks would require more work FYUI - if even possible. As I've said the limit is 4GBs and even that takes an 0x10000000 bytes hit.
So 4Gb PC2-4200-R modules do occasionally get detected via dev /memory \n .properties, however it seems unstable so far (often telephone beep) and they seem very picky about how DIMMs are populated and relocating them results in the telephone beep. Individually, they get detected without problem. It's most probably a matter of determining the right combination. Looks like they have to be mixed with UDIMMs, I couldn't get past the telephone beep by mixing them with ECC. I haven't yet tried to allocate them, but thanks shash648!
 
  • Like
Reactions: B S Magnet
So 4Gb PC2-4200-R modules do occasionally get detected via dev /memory \n .properties, however it seems unstable so far (often telephone beep) and they seem very picky about how DIMMs are populated and relocating them results in the telephone beep. Individually, they get detected without problem. It's most probably a matter of determining the right combination. Looks like they have to be mixed with UDIMMs, I couldn't get past the telephone beep by mixing them with ECC. I haven't yet tried to allocate them, but thanks shash648!
I did not get chance over the weekend to try the ram in my Quad, this is interesting findings though. I may just give a it whirl anyway and post the results.
 
Register on MacRumors! This sidebar will go away, and you'll see fewer ads.