| Poster | Thread |
g01df1sh
|  |
How will os 4 handle this Posted on 9-Mar-2010 12:25:13
| | [ #1 ] |
|
|
 |
Super Member  |
Joined: 16-Apr-2009 Posts: 1783
From: UK | | |
|
| Hi all
Read this today just wonder if it will be a problem for OS4 as well.
http://news.bbc.co.uk/2/hi/technology/8557144.stm _________________ A1200 ACA1232 128MB Indivison MkIICr Elbox empty Power Tower RPi3 Emulating C64 ZX Atari PS BBC Wii with Amiga emulation Vampire v4 SA |
|
| Status: Offline |
|
|
tiffers
|  |
Re: How will os 4 handle this Posted on 9-Mar-2010 13:08:28
| | [ #2 ] |
|
|
 |
Regular Member  |
Joined: 7-Jun-2007 Posts: 349
From: Perth, Western Australia | | |
|
| @g01df1sh
I'm no expert, so I can't give authoritative answers, but it sounds like some FUD mixed in there. I'm sure operating systems have been able to use block sizes larger than 512 bytes for some years now (I'm sure win2k allows you to choose block size larger than 512bytes, although maybe not up to 4096 bytes) Looking at wikipedia this may actually be cluster size.
I'm farily sure that even HDToolBox lets you specify the block size when creating your partition. I've seen posts for SFS and PFS specs saying they can address filesystems up over 1TB using block sizes larger than 512 bytes Example.
I guess the HDD manufacturers don't want to appear like they didn't make it clear enough, and saying "anything older than a currently supported release of your OS will probably fail." Maybe the OS manufacturers have chimed in as they can see that this could boost the slow sales of their current OS with our current economy.
I also question the figures of 5ms and 10% performance reduction. I mean come on! Drives have seek times around 5ms these days don't they? How can a dedicated processor add anything like a 5ms delay, unless by design? 10% performance reduction also sounds FUDdy to me. 10% of 3Gb/s (i know it's just burst) or 300Mbps is a LOT to lose for such a simple hack.
Can anyone state with any certainty whether the figures are even close to accurate?
tiffers Last edited by tiffers on 09-Mar-2010 at 01:29 PM. Last edited by tiffers on 09-Mar-2010 at 01:20 PM. Last edited by tiffers on 09-Mar-2010 at 01:15 PM.
|
|
| Status: Offline |
|
|
NutsAboutAmiga
|  |
Re: How will os 4 handle this Posted on 9-Mar-2010 13:14:28
| | [ #3 ] |
|
|
 |
Elite Member  |
Joined: 9-Jun-2004 Posts: 13047
From: Norway | | |
|
| |
| Status: Offline |
|
|
amije
|  |
Re: How will os 4 handle this Posted on 9-Mar-2010 13:36:46
| | [ #4 ] |
|
|
 |
Regular Member  |
Joined: 28-Apr-2006 Posts: 401
From: Thessaloniki Greece | | |
|
| @g01df1sh
nonsense. from the era of win2000 ntfs has 4k as default. who they try to fool? looks like another ''buy win7'' trick from ms to me.
btw afaik sfs works best in 512. _________________ A3040 AmigaOS3.9 A1200 060/BPPC AmigaOS3.9/4.0 Sam440ep AmigaOS4.1.2 Pegasos2/G3 AmigaOS4.1.2/Morphos2.7/Debian6.0/OpenSuse11.1 MacMini/G4 1.5 MorphOS 2.7/OSX 10.5.8
Long Live Amiga  |
|
| Status: Offline |
|
|
A1200
 |  |
Re: How will os 4 handle this Posted on 9-Mar-2010 13:51:05
| | [ #5 ] |
|
|
 |
Elite Member  |
Joined: 5-May-2003 Posts: 3130
From: Westhall, UK | | |
|
| Aside there could be (is?) support on XP for 4k blocks, people who have the latest hardware genrally have the lastest software anyway. You wouldn't expect to see a very fast new motherboard, graphics card and hard drive running windows 95, because the OS wouldn't best perform, or be supported at all. _________________ Amiga A1200, 3.1 ROMs, Blizzard 1230 MKIV 64MB & FPU, 4GB DoM SSD, Workbench 3.1 |
|
| Status: Offline |
|
|
ChrisH
 |  |
Re: How will os 4 handle this Posted on 9-Mar-2010 14:02:12
| | [ #6 ] |
|
|
 |
Elite Member  |
Joined: 30-Jan-2005 Posts: 6679
From: Unknown | | |
|
| I guess this could be a reason to switch from SFS to JXFS, since I presume the latter is more likely to be quickly updated to handle Aligned 4K blocks.
@g01df1sh If the OS *writes* less than a 4KB block (say 0.5KB), then the harddrive will have to emulate this by first *reading* a whole 4KB block, modifying the relevant part (0.5KB), anf then writing the whole 4KB block back. i.e. a single 0.5KB write could turn into a 4KB read + write. (Just remember that the time taken to write 4KB is barely longer than the time taken to write 0.5KB - the real problem is a read being required for every write.)
And even if you use 4KB blocks, if the OS doesn't ensure they are aligned to 4KB, then the harddrive will have to emulate this by accessing *two* 4KB blocks. Therefore a 4KB read could turn into an 8KB read.... while a 4KB write could turn into an 8KB read PLUS write. _________________
|
|
| Status: Offline |
|
|
NutsAboutAmiga
|  |
Re: How will os 4 handle this Posted on 9-Mar-2010 14:03:02
| | [ #7 ] |
|
|
 |
Elite Member  |
Joined: 9-Jun-2004 Posts: 13047
From: Norway | | |
|
| |
| Status: Offline |
|
|
NutsAboutAmiga
|  |
Re: How will os 4 handle this Posted on 9-Mar-2010 14:06:44
| | [ #8 ] |
|
|
 |
Elite Member  |
Joined: 9-Jun-2004 Posts: 13047
From: Norway | | |
|
| @ChrisH
Or if it buffer up the 4KB, so that first read of 0.5Kb is 1ms slow all other reads of 0.5Kb turn in to 0.1us delays.
While I agree that disk writes will be slow, because, if 4K buffer is not filled whit right 0.5K blocks it has to be first read from disk, then buffer has to be modified whit 0.5Kb bock write and then written back to disk.
Last edited by NutsAboutAmiga on 09-Mar-2010 at 02:11 PM. Last edited by NutsAboutAmiga on 09-Mar-2010 at 02:10 PM.
_________________ http://lifeofliveforit.blogspot.no/ Facebook::LiveForIt Software for AmigaOS |
|
| Status: Offline |
|
|
KimmoK
|  |
Re: How will os 4 handle this Posted on 9-Mar-2010 14:11:45
| | [ #9 ] |
|
|
 |
Elite Member  |
Joined: 14-Mar-2003 Posts: 5214
From: Ylikiiminki, Finland | | |
|
| |
| Status: Offline |
|
|
NutsAboutAmiga
|  |
Re: How will os 4 handle this Posted on 9-Mar-2010 14:22:40
| | [ #10 ] |
|
|
 |
Elite Member  |
Joined: 9-Jun-2004 Posts: 13047
From: Norway | | |
|
| @KimmoK
Slow is definition of some that not the fastest.
While Amiga where not able to keep up whit graphics in 1993, when Mac where released whit modern graphic cards and PC starting to be modern.
If time line where different, if only AGA where released in 1990, not 1992, and Amiga4000 where shipped whit S3 Virge (aka Cybergraphics gfx card) things might been different for CBM Amiga computers back in 1994. Last edited by NutsAboutAmiga on 09-Mar-2010 at 02:23 PM.
_________________ http://lifeofliveforit.blogspot.no/ Facebook::LiveForIt Software for AmigaOS |
|
| Status: Offline |
|
|
KimmoK
|  |
Re: How will os 4 handle this Posted on 9-Mar-2010 16:14:51
| | [ #11 ] |
|
|
 |
Elite Member  |
Joined: 14-Mar-2003 Posts: 5214
From: Ylikiiminki, Finland | | |
|
| @NutsAboutAmiga
IIRC, Amiga did not yet drop behind the others in GFX in 1993. UPDATE around 1993/1994: Quadra 605 GFX: - 512kb...1024kb - 832x624 high color mode - max 1152x870 256 colors Centris 650: "The higher-end model also came with 1MB VRAM installed, enabling 16-bit color at 640x480 resolution. The availability of 16-bit color was significant, as it was the standard bit depth of Apple's then-new Quicktime video standard. " A4000 had 2Mb videoRAM, and up to 1440x780x262144/24bit colors in PAL mode & being the only standard HW capable to play full screen video, etc.. more from 1993: http://www.macmothership.com/gallery/newads2/Be-At-Home-2.jpg
Neither did 3D hardware acceleration become widely used until late 90's.
But a lot more effort should have been put to the marketing and a little bit more in the output of new products. And less on the CBM PC.
Fo the appearance of 3D games, it might have been enough if every Amiga would have shipped with both chip and fast RAM and OS had supported RTG.
It would have been interesting to see if CBM had managed to start using third party GFX chips early enough, instead of the inhouse ones. Last edited by KimmoK on 09-Mar-2010 at 05:35 PM. Last edited by KimmoK on 09-Mar-2010 at 04:46 PM.
_________________ - KimmoK // For freedom, for honor, for AMIGA // // Thing that I should find more time for: CC64 - 64bit Community Computer? |
|
| Status: Offline |
|
|
asymetrix
|  |
Re: How will os 4 handle this Posted on 9-Mar-2010 16:21:02
| | [ #12 ] |
|
|
 |
Cult Member  |
Joined: 9-Mar-2003 Posts: 868
From: United Kingdom | | |
|
| |
| Status: Offline |
|
|
Jolo
|  |
Re: How will os 4 handle this Posted on 9-Mar-2010 17:14:24
| | [ #13 ] |
|
|
 |
Member  |
Joined: 7-Oct-2006 Posts: 30
From: Unknown | | |
|
| You must distinguish between a file system block size and a hardware logic block size (sector size). Cluster sizes and file system block sizes can be multiple of 512 bytes, and that's what they are today already. The new hardware logic that gets introduced in January 2011 for hard disks will always use a 4k sector size, not anymore 512 bytes, but to remain compatible, the interface will still be accessed as if a sector size of 512 bytes would be used. So, the good thing is that even old file systems can deal with new hard disk drives. The bad thing is that such old file systems, which are not able to align write processes to boundaries being octuple of 512 bytes, will be slower with such new hard disks than with hard disks that still use a genuine sector size of 512 bytes.
All in all, nothing to worry about.
By the way, Western Digital already ships such hard drives.
|
|
| Status: Offline |
|
|
Leo
|  |
Re: How will os 4 handle this Posted on 9-Mar-2010 19:14:06
| | [ #14 ] |
|
|
 |
Super Member  |
Joined: 21-Aug-2003 Posts: 1597
From: Unknown | | |
|
| Quote:
IIRC, Amiga did not yet drop behind the others in GFX in 1993.
|
Of course it did.
AGA was unsuable in anything higher than 320x256 in 256 colours... Even lowres was slow in 8bit. Nothing comparable with the Mac of that time for sure._________________ http://www.warpdesign.fr/ |
|
| Status: Offline |
|
|