76

(73 replies, posted in Vista / 7 DriverPack Mass Storage)

IaStor (Intel matrix storage) is up to 8.6 WHQL now...

77

(88 replies, posted in DriverPack LAN)

It's now posted on Intel but the download link is broken. What a bunch of incompetent buffoons!

OK test results with 23 driverpacks (latest nightlies + all 3rd party packs except for HID and printers):

Traditional presetup.cmd: 5 minutes 20 seconds
Modified presetup.cmd: 3 minutes 45 seconds.

This is on a laptop drive. Not a speed demon. Not as much as I would hope, but still cuts time by a large margin smile

Notes: There seemed to be a long pause to extract packs no matter the size, even the tiny 3rd party stuff. Wonder what that is? It's definitely not the CD/DVD drive, that is observably no longer touched with this new test method.

Also: the extraction routine will cause low memory warnings on 512MB RAM machines or less! It still finishes unattended but perhaps something as robust as this method should have a minimum requirement of a 1GB RAM PC?

79

(34 replies, posted in News)

It would be nice to have some more thorough documentation for the Tool, packs, updatepacks, etc. More than what a readme.txt provides.

80

(107 replies, posted in 3rd Party DriverPacks)

I alway thought that you needed vendor specific drivers for the reasons you state above - however my G9 uses the generic and all the default buttons work with generic, plus the dpi adjustment - and get this, with the latest mouseware (at least for the G9), the default install routine uses the generic MS driver! You have to manually select to insist on using the logitech driver.

Doing that now smile

82

(88 replies, posted in DriverPack LAN)

v13.5 out, not official (yet). For those who like to mod ahead of the nightlies smile

http://www.station-drivers.com/page/intel%20lan-wan.htm

For those who like to mod ahead of nightlies; 9.1.0.1012 WHQL is out.

http://files.laptopvideo2go.com/chipset … 0.1012.exe

Wont update the OP or title since these are nowhere to be found on Intel's site yet smile

Like I posted earlier, you do not need a RAID 0 array to see the speed benefits of copying to the HD first.

My test w/ DriverPack Graphics A was on a laptop 7200rpm drive, which by many standards is slow compared to a nominal desktop drive.

85

(107 replies, posted in 3rd Party DriverPacks)

I wonder if it's really worth it to have a HID pack that covers mice that otherwise can work fine with MS supplied generic HID drivers? The touchpad stuff is definitely useful, but what a pain dealing with yellow devices on simple mice...

86

(34 replies, posted in News)

I'm in. At home I've been 100% vista 64bit. I've had some quirks with the available integration tools + WAIK so this is something I want to tackle.

87

(107 replies, posted in 3rd Party DriverPacks)

Fragbert wrote:

Logitech G5 Laser gets the yellow exclamation point in device manager with current HID pack. Noticed the driver provided is 4.60, setpoint for this mouse is now at v4.70. I'll try the new driver to see if that fixes.

=============
Input Devices
=============
HID\VID_045E&PID_00B4&MI_00\7&2E835FE5&0&0000               : HID Keyboard Device
HID\VID_045E&PID_00B4&MI_01&COL01\7&AAC2223&0&0000          : HID-compliant consumer control device
HID\VID_045E&PID_00B4&MI_01&COL02\7&AAC2223&0&0001          : HID-compliant device
HID\VID_046D&PID_C049&MI_00\7&1693DD53&0&0000               : Logitech HID-compliant G5 Laser Mouse
HID\VID_046D&PID_C049&MI_01&COL01\7&363C115&0&0000          : HID-compliant device
HID\VID_046D&PID_C049&MI_01&COL02\7&363C115&0&0001          : HID-compliant device
6 matching device(s) found.

Update: v4.70 fixes issue for the G5 mouse, but dell branded logitech mice now go yellow. Logitech hardware seem very picky on drivers in this pack.

88

(107 replies, posted in 3rd Party DriverPacks)

Logitech G5 Laser gets the yellow exclamation point in device manager with current HID pack. Noticed the driver provided is 4.60, setpoint for this mouse is now at v4.70. I'll try the new driver to see if that fixes.

=============
Input Devices
=============
HID\VID_045E&PID_00B4&MI_00\7&2E835FE5&0&0000               : HID Keyboard Device
HID\VID_045E&PID_00B4&MI_01&COL01\7&AAC2223&0&0000          : HID-compliant consumer control device
HID\VID_045E&PID_00B4&MI_01&COL02\7&AAC2223&0&0001          : HID-compliant device
HID\VID_046D&PID_C049&MI_00\7&1693DD53&0&0000               : Logitech HID-compliant G5 Laser Mouse
HID\VID_046D&PID_C049&MI_01&COL01\7&363C115&0&0000          : HID-compliant device
HID\VID_046D&PID_C049&MI_01&COL02\7&363C115&0&0001          : HID-compliant device
6 matching device(s) found.

chud wrote:

Fastest (should/would) be extracting straight from CD/DVD from packs packed with minimal compression to HDD. Reading a sequentially large file is quick (think ripping an ISO, or verifying a burn).

No, just trust me on that one. If you copy the archive to disk, then you are just reading from CD (decent speed). Extracting directly off CD means you will be checking back and forth to the archive after each iddy-bitty file (horribly slow due to CD seeks). You would think that the archive would be loaded/extracted from memory but in practice it just doesn't work that way.


chud wrote:

HD>HD extraction is going to be slower unless you have a super fast PC/HDD.

Actually, a current generation 7200rpm drive is all it takes to cut your extraction time in half, see my above extraction time test smile

chud wrote:

If anyone is interested I make a DVD with the added packs, create ~3.5Gb padfile, add it with UltraISO setting pad file priority to 9 (means it gets put on centre of disk, low LBA) - all the useful files are then placed on the outside of the DVD so xfer rates are about 4x faster overall. You can also keep the DP folder priority 0, and everything else 1 if you want max extract speed (this will place them on outside of disk, so you get about 22x read over 4x on inside)

Actually I experimented with this before too - depending upon the PC, the overhead on extraction based on compression can be minor on a decent PC. And no doubt you will be adding time to copy from CD/DVD if your uncompressed file is 3.5GB vs 350MB compressed.

Basically you want these files to be put on the fastest medium (HDD) as quickly as possible (OEM setup) before they do any work. Take the CD/DVD out of the picture ASAP.

Let me make it very clear - you absolutely, positively, unequivocally, unconditionally MUST copy any archive to the hard drive first before extracting (high or low compression). You DO NOT WANT to ever start extracting an archive full of millions of little files DIRECTLY off a CD. Your grandchildren will be gray by then smile

OverFlow wrote:

still thinking... so a M3 slipstream might be like this...

base then repacks the users selected packs into a single large archive with minimal or no compression.
at this point do we still copy to the HDD first or does it not matter without the overhead of compression?

all combined into a single pack no compression
and with and without precopy to HDD...

OverFlow wrote:

I agree

enter the to be developed Method Three (M3) and SoFI
with huge amounts of space that are available for install media it is my belief that compression is not needed at all.

SoFI (SAD) can be called from runonce and will run against a M1 or a raw extracted folder source.


or the M3 i have been thinking about for a while
IE
extract all DriverPacks to D\
copy D\ from your install media to SystemRoot from presetup.cmd
instead of the extraction routine that is there.
I was not sure that m3 would be faster because of the overhead assosiated with the thousands of small files.
it does create a searchable ODD for the 'found new hardware wizard' as a side effect and could be used as a SAD.

this is the outline for M3...

Yikes that will be just as slow. Your goal is to overcome the SUCK that is DVD/CD media. The best method is to copy a few large files (driverpacks still compressed) to the HD, not a million little files. That will slow the CD/DVD to a crawl. DVD/CDs do OK with large sequential file copies, nothing else. That is why a modified method 2 will be ideal - copy the large compressed DP's first, then uncompress once on the HD.

You are right about the compression not being needed, but keeping all those millions of files as one 7z IS needed, so perhaps have the DriverPacks repacked as simple store or zip with little to no compression. That will make things quicker with there being no overhead to uncompress the DriverPacks.

Extracting zip files, let alone ultra-compressed 7z's, directly from DVD/CD can be much slower than first copying the compressed file to HD then uncompressing. The overhead from the horrid seek times is the wrench in the wheels I think. Copying files over in a sequential file transfer, especially large files, from disc does not take as big a hit.

Try it yourself with DriverPack Graphics A smile

On my rig, it takes approximately 41 seconds for DriverPack Graphics A to decompress directly from DVD to hard drive.
Copying DriverPack Graphics A first from DVD to disc takes a mere 4 seconds, and then only 20 seconds to decompress. The time has almost been cut in HALF!!

DVD/CD is still a profoundly slow medium.

I notice that during the "fake setup" in method 2 that the driverpacks (with shortened names) are being extracted directly from the CD/DVD which slows this process considerably. Would it not be faster if the contents of the OEM folder were copied first to the HD before the fake setup / extraction starts?

mr_smartepants wrote:

I'll look into the other Intel drivers.  Maybe I just have to rearrange the layout of the other folders to eliminate the overlap.

I might be mistaken in thinking mass_storage for intel was more complex. On the surface, DP MS is only three-tiered to segment the intel generations. DP_G_B has four already?

mr_smartepants wrote:

I'm not going to update I1\ to Intel GMA driver v14.37.0.5009 because they dropped half the HWIDs and added nothing.
That's just dumb!

?

It's the latest driver with a litany of bug fixes to the hardware supported by it. I thought all their new releases since this year were 3 and 4 generation exclusive anyway? http://downloadmirror.intel.com/17178/e … xp_gfx.htm

If your above reasoning is why you wont update Graphics B, you may as well just abandon it then, because you wont be seeing any of those HWIDs back in future releases either, nor any new hardware support till Q2. I think driver updates that fix things are just as important to this project as driver updates that add support to new hardware. It's a royal pain to support older hardware in DriverPacks, just like with Intel Matrix storage, but worth it IMO. Intel is making a serious effort to make their IGPs no longer a laughing stock and if you follow their developer forums, their driver releases continue to fix huge game/app bugs, and this release is no different.

I've found that Dell PE's are very sensitive to drivers outside their own releases, particularly intel graphics IGP (GMA) and broadcom NICs.

I have no clue why Dell ditched integrated Intel and went with broadcom netxtreme for their servers. Intel NICs have been, are, and will be the best server NIC out there (at least in the 10/100/1000 ethernet realm). Broadcom are crap interfaces and adapter teaming is still a gamble with their latest driver releases.

mr_smartepants wrote:

I don't see it.  They've still got 14.36.4.5002 listed.

Follow the link at the top. I always edit the link and title when a new release is out.

v14.37.0.5009 out

99

(88 replies, posted in DriverPack LAN)

newsposter wrote:

You're slowing down fragbert, those drivers have been up for a week now............

I've been meddling in Vista x64 most of this month smile

100

(88 replies, posted in DriverPack LAN)

Eh..HEM...no stationdrivers link needed, Intel has 13.4 up on their own site, first post link updated smile