Classic Mac Hardware (Troubleshooting, Upgrading, & Modifying) > Storage Technologies

64 bits PCI on G4 working and confirmed!

<< < (10/10)

macStuff:

--- Quote from: stephenvalente on February 02, 2017, 02:03:25 AM ---
--- Quote from: MacTron on February 01, 2017, 09:06:52 AM ---The G4 MDD 1.42 can be overclocked to 1.58-1.66 Ghz or upgraded with a  Sonnet MDX duet 1.8 ghz.

--- End quote ---

Not too fussed about overclocking. More interested in finding a SATA card, with or without RAID. Read a few of the threads around, and it seems a difficult thing to obtain.

--- End quote ---

something that alot of regulars on this site are seemingly oblivious to; i keep posting trying to draw peoples attention to the fact that a modern solution needs to be worked out but im getting very little cooperation or assistance with chasing that goal. everyone is asleep at the wheel around here. there are people posting sonnet tempo sata pci cards for 400$ on ebay because there is no other solution available to anyone untill some real progress is made with a fix for this.


--- Quote from: stephenvalente on February 03, 2017, 06:59:39 AM ---IOCrest SATA II 4 x PCI RAID Host Controller Card SY-PCI40010

https://www.amazon.com/gp/aw/d/B002R0DZZ8/ref=cm_cr_arp_mb_bdcrb_top?ie=UTF8

This card is said to work in a Power Mac G4 by a reviewer, but non-bootable? Anyone else used it? I'd happily use this alongside a small SSD with an IDE-SATA adaptor card for a solution to speeding the storage side of things up.

--- End quote ---

i think you misread - that was someone using an HP server that used the name G5; (from what i can see) that is a PC; also; that card is only 32bit / it would be much more gainful to focus on testing a 64bit card such as the sil3124; specifically the Silicon Image PB3124 that i have been posting about; which already contains a Macronix 040 flash chip in a socket; which is a verified chip usable by the 1S2 Firmtek firmware;

mePy2:

--- Quote from: MacTron on January 29, 2015, 09:48:28 AM ---When a Hard Disk data transfer occurs, alongside the user data, some other data is transferred also (control data, address data, etc ) this usually use around 20% of bandwidth. That's why if we measure actual speed of file transfer over a ATA 100 the most we can expect is 80% of the ATA 100 speed (100MB/s), ie 80 MB per second.

I've used a SATA3 Samsung 840 EVO (Max Read = 540 MB/s Max Write = 410 MB/s)

http://www.samsung.com/uk/consumer/memory-cards-hdd-odd/ssd/840-evo/MZ-7TE120BW

In to some of my test, to avoid the limiting factor of the storage device.

But  according to this chart:



190 MB/s on raid mode is less than expected (230-244 MB/s) on the Seritek 64 bits card. Including the 20% of bandwith lost.

So what's the limiting factor now? And what we can learn on this?
The 410-540 MB/s of the SSD sure is not.
The 256 MB/s of the 64 bits PCI may be not.

In Raid Mode we surpass the 150MB/s of the SATA 1 speed. ie A raid of 2 HD (150+150 = 300 Mb/s) surpass the 64 bits PCI also.
Will the Seriteck chip capacity the limiting factor? probably yes.
Two Seriteck cards with a SSD each, in RAID mode would reach the MAX 64bits PCI bus (256 MB/s)?
Who have two 64bits Seritek cards?

A 64 bits data can be divided in high (32bits) and low (32bits) part. Can the 64 bits bus use the full bandwith (32+32 = 64 bits) even with 32 bits cards?
If this were true: Three Sil3112 with a SSD each, in RAID mode would reach the MAX 64bits PCI bus (256 MB/s)?

--- End quote ---

Interested in this.
May you all come discuss about it here? https://discord.gg/8WHpu2m

Navigation

[0] Message Index

[*] Previous page

Go to full version