DIY eGPU experiences

It's pretty damn amazing how far you guys have come with these projects. *Kudos*

 

I was asked to bench my G73 for this thread.

I will add my results however it must be stated for my conscious to be clear that the benchmarks being used are ridiculous.

First of all 3dmark06 is too old now, it is easy cpu bottlenecked by any modern high end and even upper mid ranged video cards. Its also not going to play well with a quad core like whats in the G73 so I will hit a cpu limit way way way before I hit a gpu limit so you will be looking at the score thinking "wow thats low" but you wont be looking at the true score for the gpu and that is what this thread is all about.

Likewise the RE5 benchmark has a variable and a fixed. Fixed is the same in every run this is the one you would want to use to compare multiple systems, while variable is... well variable.

It lets AI and other components randomize the benchmark so the scores can fluctuate between each run. But by far the largest problem is the requested resolution for this test 1200x800? who games at that resolution?

Again this is going to be incredibly easy for the gpu and cause a massive cpu bottleneck and give invalid scores and false representations of a high end system.

In 2006 these would have been ok but now in 2010 technology has moved way beyond these low standards.

If you want to do this and do it right bench something like 3dmark09 Vantage, and then post ONLY the gpu score rather than the combined score so your comparing the gpu power and not the cpu + gpu power.

Just had to say that because too many ill informed eyes may look upon the information presented here and not understand what they are really looking at.

 

G73 w/ 720 gets ~12400 3dmarks06 with 10.3 cat
can't do a comparison between 5870m and 5750 cuz there is no expresscard in it....

 

x1E/x2E tweak mode on Intel chipsets gain 15-30% performance on odd numbered ports using ATI cards or older NVidia cards (GTX4xx is not affected)

NOTICE: ATI or older NVidia cards using Intel notebook chipsets' mPCIe/expresscard x1 1.0 are underperforming by 15-30%. NVidia's latest GTX4xx cards are not affected as shown here. Affected Systems: Intel ICH7/8/9M I/O-equipped chipsets below. Video cards tested: ATi HD5750 below and nVidia 9800GT. GTX4xx cards are NOT affected nor are AMD/NVidia chipsets so this problem is because of ATI/NVidia x1 1.0 pci-e handshaking implementation. series-5 chipsets do not see any performance improvement with the tweak below as documented in the DV4-2000+HD5750 implementation. RECOMMEND: escalate the issue to ATI/NVidia to get a solution. The problem has also been logged on Intel's community support page here and ATI's forum here.

Using ATI cards or older NVidia cards on Intel I/O chipsets' x1 1.0 pci-e ports has handicapped performance requiring setting it to x2 width to get a 15-30% performance improvement. This workaround applies only if the port used for the eGPU occupies an odd numbered port. It would appear that the these older ATI/NVidia cards may be running half-duplex, requiring setting it to x2 width to get the full x1 1.0 bandwidth.

The Nvidia MCP79 chipset based Dell XPS M1340 benchmarks here blitzes all the Intel machines at x1 1.0 speed. It gets full x1 1.0 bandwidth.

The tweak discovered to enable full x1 1.0 speed on a Intel pci-e port requires the mPCIe or expresscard to be an odd numbered port with no device on the next sequential port, eg: port1/port3/port5 with no device on port2/port4/port6. I'm calling this x1E mode, the "E" standing for Intel-Enhanced.

x2E being similar but requires a port1+port2 setup, with either lane3 or lane4 or both having NO physical device attached to them. port1 is set to x4 mode, running with only the first 2 lanes. x2E also offering 15-30% faster performance than x2 alone.

NOTE: if x1E is enabled on port3 or port5, the effect is the same as when x2 is engaged on them - port1 or port3 respectively will also be switched to x2 which in turn render the port2 or/and port4 non-operational.

Background

With the arrival of a PE4H V2 I attempted a x2/x4 link with the DV4, port layout here, hoping that Intel had made a mistake somewhere and I could link say port1+port3 as a dumbed down x4 link working in x2 1.0 mode. It didn't work.

During the testing however a very unusual performance increase was observed. I had set port1 into x2 width and was seeing ~20% better performance. That is *without* connecting a second lane. The 3dmark06 values don't look much different, but delve deeper and see the DV4's SM3.0 value of 3dmark06 has increased from 4344 to 5247 and is *higher* than the i7-620M+HD5750 x1 1.0 implementation's 5085. SM3.0 correlates directly to 3D gaming performance. The RE5 benchmarks improve considerably, dmcv4 dx9 sees a 30% improvement!! In case you were wondering, does setting port1 to be an x4 port with only the single lane connected see any further performance? The answer is no.. it does not.

x2 versus x1E versus x1 link speed, 2GB HD4870 versus 1GB HD5750

See also x1E versus x1 benchmarking and here and here (nVidia).

System	RAM	GPU	3dmark		RE5 (*=dx10)		dmcv4.scene4		PCIScript^2	Ports	OS
			06	vant.g	var	fixed	dx9	dx10
T2050-1.6 HP DV2000^1	3.0	HD5750@x2	6735	5579 +	38.5*	18.3*	60.1	55.4	Setup 1.x	ICH7M	Win7/32
		HD4870.2GB@x2	6710	5499	40.2*	18.1*	58.7	51.9		ICH7M
P8600-2.4 HP Pav_DV4	4.0	HD5750@x1E	10151 &	5215	64.0 &	35.0 &	66.3 &	41.8	-	ICH9M	Win7/64RC1
T6600-2.2 Dell Insp_1440	4.0		9551 &	5004	60.9 &	30.8 &	66.7 &	42.1 &	-	ICH9M	Win7/32
		HD4870.2GB@x1E	9188	5486	57.0	31.2	51.4	43.6		-
T2050-1.6 HP DV2000^1	3.0	HD5750@x1E	6694	4903 +	34.6*	17.0*	59.9	43.7	-	ICH7M	Win7/32
P8600-2.4 HP Pav_DV4	4.0	HD5750@x1	9461 &	4805	52.8 &	26.0 &	41.7 &	36.4 &	-	ICH9M	Win7/64RC1
T6600-2.2 Dell Insp_1440	4.0		8997 &	4432	51.6 &	25.7 &	44.7 &	38.0 &	-	ICH9M	Win7/32
		HD4870.2GB@x1	7913	4576	42.4	22.7	34.9	32.1	-	-
T2050-1.6 HP DV2000^1	3.0	HD5750@x1	6560	4219 +	33.2*	16.2*	43.8	36.7	-	ICH7M	Win7/32

&=overclocked HD5750 from 700/1150 to 790/1430
^1 port3 required mPCIe socket to be soldered onto systemboard's pads marked MINI2
^2 only first instance of PCIScript shown, duplicates per system have been removed.
Bold - benchmarking showing x2's superior performance even with a greatly inferior CPU

About HD5750 versus HD4870

We see the HD5xxx series has made more inroads at minimising pci-e bandwidth since the HD5750 outperforms the HD4870 in virtually all cases, particularly on low x1 1.0 bandwith setups. This contradicts what we'd expect. x16 2.0 benchmarks results indicate a HD4870 to be onpar to a HD5770 as shown here.

It is not clear then if 2GB VRAM does have any bandwidth sparing effects as I have suggested in my posts. Would need a direct comparison of the same series card (eg: 1GB HD4870 versus 2GB HD4870 or 1GB HD5850 versus 2GB HD5850) to see how much extra performance 2GB VRAM provides.

Overall I would advise against a HD4xxx card in favor of a more pci-e bandwidth efficient HD5xxx card.

How can I switch my port into x1E mode?

This requires an accessible odd numbered mPCIe or expresscard port with no device connected to a next even port as otherwise it won't work [port1 for DV4, port5 for Inspiron 1440]. If you have limited options, with perhaps a LAN card or some other device occupying a next even port, consider finding the datasheet to figure how to disable it. Or else if can identify Vcc power pin to the chip, it could be possible to do a cut the pin to disconnect it from being on the PCI bus. I really wanted to see if x1E could work on port3 of the DV4 since it's the more convenient expresscard slot, but the LAN card on port4 prevented that. I attempted to write 00 data to the LAN card's PCI space to try to disable it the device was still active. So x1E is only possible on the mPCIe port1 on the DV4.

The no-brainer solution is to use the DIY eGPU Setup USB Bootdisk to switch your port to x2 via the menus, or using 'iport x2 [port]' at the commandline.

Spoiler below details what 'iport x2' does which can be implemented using the older DIY ViDock bootdisk or, if you really want to not use the 'iport' command then can do this manual way using the new DIY eGPU Setup 1.x

Spoiler

1. Identify RCBA address of your system using . Disregard the trailing 1 treating it as zero. Eg: if setpci returns 0xFED18001 then RCBA is 0xFED18000. This can be done using Baredit in Windows as shown here. 2. If your bios has a next even *enabled* port, (eg: port2/4/6), do steps 3-4 otherwise go to step 5.
3. Add RCBA+0x3418 (that's hex.. can use calc.exe in scientific mode). That's were the function disable register is. 4. Set bit 17/19/21 to 1 to disable port 2/4/6 as visually shown here. 5. Add RCBA+0x224 (that's hex.. can use calc.exe in scientific mode). That's were the port configuration is stored. 6. Add appropriate details below to your DIY ViDock bootdisk PCIScript.
Code: :: Disable next even port port2/port4/port6 if bios has it enabeld pt MEM write 4 [RCBA+3418h] [change bit 17/19/21 to 1] ::choose port1 (+ non occupied port2) to be x2 high priority port pt MEM write 4 [RCBA+224h] 0x81 ::choose port3 (+ non occupied port4) to be x2 high priority port ::This also sets port1 to be a x2 port [ no way to prevent this! ] pt MEM write 4 [RCBA+224h] 0xA2 ::choose port5 (+ non occupied port6) to be x2 high priority port pt MEM write 4 [RCBA+224h] 0xC4 ::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::: :: Examples: create a x1E link. A 20% faster x1.0 link :::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::: :: example #1: HP_Pavilion DV4 using mPCIe port 1 :: 'setpci -s 0:1f.0 f0.l' returns RCBA=0xFED1C001 pt MEM write 4 0xFED1c224 0x81 :: example #2: HP_Pavilion DV2000 using mPCIe port 3 :: 'setpci -s 0:1f.0 f0.l' returns RCBA=0xFED1C001 :: Disable port2 (as port1 becomes x2 port) and port4 pt MEM write 4 0xFED1f418 0x3A0061 pt MEM write 4 0xFED1c224 0xA2 :: example #3: Dell Inspiron 1440 using expresscard port5 :: 'setpci -s 0:1f.0 f0.l' returns RCBA=0xFED18001 pt MEM write 4 0xFED18224 0xC4

What about users that have a PNP system without needing a DIY ViDock bootdisk?

This could only apply if you intend to setup port3 or port5 as a x1E link. port1 is the "root port". Disabling it affects all other devices off port1-port6, so it's not a good idea to use this method on port1. Use the DIY ViDock bootdisk instead to set a x1E port1.

The following worked on the Inspiron 1440 but caused a freeze on the DV4. The DV4 is using the off-limits port1.

1. Disable the port you want to set as x1E and any associated even ports as explained in steps 2-4 above using Baredit like shown here
2. Do a device manager scan or use ' devcon rescan'
3. Change desired port into x2 width by writing one of the following to RCBA+224h using Baredit.: 0x81(port1), 0xA2(port3) or 0xC4(port5).
4. Connect your EC2C or PM3N and power your video card.
5. Enable the port by reversing step 1. When you click "write" in baredit the port is enabled, now in x1E mode.
6. If Win7 doesn't automatically detect the port, do a device manager scan or do ' devcon rescan'

Only this sequence of steps resulted in a stable setup. The sequence above can be scripted. The baredit command can be 'saved', which saves it as \bar_save\bar_edit.ini. Running bar-edit from that directory then executes the commands in bar_edit.ini without user intervention.

Quirks

A standby/resume of the Inspiron 1440 resulted in port5 reverting back to x1 width. DV4 maintained the port state after a standby/resume. Suggest check your port with Everest after a resume to see if your bios is quirked this way.

 

Updates of my VPC-Z11 Core i7 notebook.

The new RE5 benchmark is in the following link.
http://zwhraq.bay.livefilestore.com...vYL3WrJeqRyVufK8dVE/RE5DX9_100414_1458212.BMP

I can't use Ati Control Center as it keeps complaining no video card is found, other than that, it work fine, that said, no overclocking possible. I tried other overclocking tools, none of them detects ATI 5750!!

 

Shouldn't your onboard videocard be on port 2, northbridge?

Did you try Ati tray tools?

 

i have a question. I remember in the first page it was mentioned that you configure the port 1 for x4 mode right. Why can't you do this with the pe4h to get a x4 speed on one port instead of ganging up two port to get x2

 

nando - where or how did you get the pe4h v2? currently only v1.0 is orderable at hwtools.net

 

Does it means a possibility , if I have port #1 not used, I can disable it to speed-up ExpressCard's port#4?

 

20% speedup is applied to odd number ports(1,3,5...) connected to external video card. the other needed condition is just one adjacent port(2,4,6,..)be empty.
so, you cannot speed up port 4.

and.. i think the disabling port #1 would result in disabling all pci express ports. because port #1 is a root port of pci express ports.

...

 

Ok. And how about High Priority Port? I suspect, that some ports are slower because energy saving. Maybe HPE port have full bandwidth all the time?

 

hi nando4.
first of all i appreciate you for this new post with a solution to my headaching analyzing problem on my machine.

20% speedup's applied well to ICH7(945 chipset family).
refer to attached files.

...

so.. i've got a question in your coding. is it all right in above coding,
like this..
pt MEM write 4 [RCBA+224h] 0x82

if x2 mode, setting high priority aside, then 0x81 isn't right?
isn't it miswritten?

 

not so big effectiveness i've found...

i suppose that is effective in the case of that interrupt crashes frequently occur to strangle the pci-e graphics connection...lol

 

You are correct. It should be 0x81. I've changed the post. Yoresults clearly showing the 20% extra performance.

If you'd like your system to be included on the first post, can you post Everest port layout for x1E mode, your script and 3dmark06 results in your above post? We have no HD5670 results so far so it would be useful. Your video card will have the designation HD5650@x1E (Intel Enhanced x1 1.0). I'll be updating the DV4 and Inspiron1440 results on the first post to be HD5750@x1E once I gather all the pieces to this.

 

Good job finding out about x1E, Nando! Since my DIY Vidock is connected on port 5 would it be possible for me to attempt to set port 5 to x2 and get that extra 20% performance? My only concern is that Everest doesn't see port 6...would this make a difference since I don't actually need to physically hook up to port 6?

 

Yes - that's what I did on the ICH9M Inspiron 1440. I set port5 as a x2 port and got my 20%. Better that port6 is not active, since really the good practice would be to disable it in the function-disable register. The Intel datasheet doesn't say you can make port5 a x2 port, it's a reserved setting. But Everest tells us: see Inspiron 1440 port5@x1E!!

Please re-do benchmarks/results/scripts for Asus F8SA/P in x1E mode here.(!!) It scales everything up a notch or two. Can see DV4 is near the top of the list now on the first post. At least until others start doing x1E mode. Eg: the Sony Z115 i7-620M+HD5750 has a port layout to do x1E. 20% would mean an increase from 68->80+FPS in RE5.

 

For anyone wondering (such as myself about a week ago) -->

Setup: Asus N61JQ (Win 7 HP 64bit) + DIY ViDock (PE4L with EC2C) + HD5670 + power adapter from my sata to usb adapter (google images "CABLEMAX USB SATA" for example)

No scripts or special boot required.

Boot + install latest ATI drivers + restart (with 5670 plugged in). The N61JQ's screen will be disabled but you can then use external monitor/s via the HD5670.

I took 2GB (of the 4GB) out. Booted into Windows normally + plugged the 5670 in and voila. No (code 12) and both HD 5730 and external HD 5670 works perfectly.

Obviously taking 2GB out is not ideal so I'm currently working with Nando to see if we can create a boot script.

Also to successfully boot via USB with N61JQ I had to:

- Download a Win95 DOS bootdisk image from allbootdisks.com
- Use WinImage to mount the image and extract to a DOS directory.
- Use HPUSBFW_2.2.3.zip to format the USB stick, pointing to the DOS directory for the system files (from Nandos Repository)
- Extract the DIYVDOCK-xxx.rar to the USB stick, ensuring you don't overwrite any existing files (autoexec.bat is OK) - also from Repository.

Hope this helps for those N61 owners that want to output 2560x1600 or simply want more extended monitors.

If we figure out how to work it with 4GB or more I'll let everyone know.

Thanks again to Nando4!

 

I'm still having trouble figuring out the memory allocation for this Dell Inspiron 1420. Here's the bpd file and the screenshots. I see that the systems resource consumes an obscene 3gb from 00100000h - DFEFFFFFh and I'm sure somewhere in there I can sneak in the ViDock. But I'm unsure what to look for exactly.

Thanks again for your help

 

Your PCI Bus space starts at E0000000 (TOLUD). Dell have set it so high so you have maximum RAM when using a 32-bit OS. I found this to be the case on the Inspiron 1440 as well, but thankfully it's 4500MHD allows IGP Reallocation.

In your case there are 2x256MB windows to work with. One window occupied by the 8400M, the other by misc devices. This is a hopeless situation with no room to work with. Potential solutions would be:

1. Apply IGP Reallocation to the 8400M to be relocated into 36-bit space. You have large Memory 120000000-317FFFFFF to do that in. Success depends on the NVidia driver being truly 36-bit compatible. ATI driver doesn't work when allocated there.

2. Ask Dell OR hack the bios so that it sets TOLUD to be C0000000. That would give you another 256MB window. If can disable the D_LCK bit in bios then have free reign to remap 32-bit memory as you like. Check using limitram.bat to see current state.

3. Use XP_drivers, disable the 8400M, hijaak it's 256MB window. You'd be giving up Aero and DX10.

4. Remove RAM so TOLUD is lower, freeing up 256MB window.

5. Upgrade to an IGP-based 4500MHD or newer system since 4500MHD driver supports relocation into 36-bit space.

 

hi, nando4 & everyone.

strange thing is here in attached shots below.
just now i checked my system with the everest in DMI term.
so, clicked [system slot] then 3 pci-e slots & 2 pci slots were appeared there. but, 3 pci-e slots ALL were empty. it was strange to me because of my using pci-e port1 connected HD5670 in that time to now.

so.. i'm attaching relevant pics for sharing the strange thing, and hoping to understand that problem.
additory shots are in-using-port1 shot, port1' physical socket number shot, and relevant ICH7 registry shot.
please, advise me. .. any body, ..here.. ?

 

There may be a few factors that allowed for higher DVMC4 performance...

1) 8gb of ram allows DMC4 to have 3.2Gbps of Ram to use
2) The ram is clocked at 800Mhz DDR2 as opposed to the popular 666mhz DDR2 that OEMs like to use.
3) I'm using AMD Fusion Utility which somehow improves performance by replacing Microsoft Referenced AMD CPU and ATI drivers with update AMD drivers after it shuts down a bunch of MS processes and then loads an updated version of the old AMD dual-core optimizer called AMD boost.

Other Tx2 News... the PCIe 2.0 script from the previous post is returning a checksum of 00 as opposed to 2=5gbit and 1=2.5... therefore the PCIe2.0 is still defaulting to a PCIe 1.0 x1 connection? or is it running at 5gbits with a Send/Rcv speed of 2.5gbps? So the assumed bit orders may need to be changed until we find the right combination?(have no idea).

Whether I run the PCIe 2.0 script or not... its still performing the same.

On a happy note (after all I am happy the darned thing works anyway) overclocking the 5770 to CLk: 950Mhz MEM: 1350 MHz... got better performance out of DVMC4 with 109fps AVG.

Another interesting discovery (probably proof the Expressport is bottlenecking at PCIe 1.0 speeds) is that regardless of the resolution and AA settings... RE5 is putting out the same framerates 42-47 fps.

Very happy note... the Cooler Master Elite mounting bracket fits the mounting bolts on the back of my HannzG HW223 monitor... as soon as I get my 1.5m pcie Cable... I'll post updated pics of the External GPU and monitor combined into one unit.

You all rock.

 

Confirmed! I managed to modify the Bootdisk script so that slot 5 is set to x2. Everest confirms that the port is set to 1.0 x2! Indeed there is a ~20% performance boost on both 3dmark06 and RE5. SM3.0 score has gone up from ~4061 to 4860 and I am getting 8-10 fps more on RE5!

I will be posting detailed benchmark results and the script later today. I've also noticed that suspending/resuming the laptop resets the port configuration to x1 and performance drops to x1 levels until I reboot the laptop. BTW Nando, regarding your instructions, I added 0x223 to my RCBA address (0xFED1C001 making it 0xFED1C224) rather than 0x224 (which would make it 0xFED1C225) otherwise it wouldn't work.

 

got it attached to the my lenovo T400 (t9400 4GB)used the gateway 6831 fx power adapter (120w)and an old 8400gs to test. win 7 suspend resume. drivers installed off windows update, screen lights up. works !!!
as simple as that.

now my only question is which card to buy ?
will the 5770 work without the 6 pin power feed ?
i understand that through x1 mode the performance gain over 5750 is negligible (or non existent), but for future proofing i would rather go with the 5770.

so if anyone can shed some light here i would be grateful
again this is an amazing thread !

 

Everest->South Bridge tells you is exactly what pci-e ports are occupied and by what. Disregard DMI data as it's an another layer use for inventory which can be inaccurate, or not populated with data, as you are seeing.

Great PNP setup with the T400. Newegg's base 1GB HD5770 costs only US$5 more than their base HD5750. It's about 8-10% faster than a HD5750 so HD5770 is best bang-per-buck atm. Only reason you'd go a HD5750 imho is if you were running a 12V/80W adapter with a PE4L.

The HD5770 will not work without power being supplied to the 6-pin connector. So you'd need to figure a way of splitting your 120W adapter. Might be easier and cheaper to just get a US$20 ATX PSU to drive it.

Excellent! 2 confirmations + my own. I'm continually updating the x1E mode documentation as reports come in. Points added about suspend/resume, how to get it back to x1E mode without a reboot using Baredit, and the extra "1" on the RCBA value.

 

I found the sourcecode for setpci ... I'm going to see if I can compile an x86_64 version for windows and run the pcie2.0x1 script in CMD and see if that makes the 5gbps transfer speeds happen. I'll update with results.

 

Hello nando4 and everyone...again,

As of now, hd5750 is running like a charm. It is amazing to see how much difference the GPU makes.

I also tried doing x1E via DIY ViDock bootdisk.
Apparently, my system's RCBA address(FED1C001) matches DV4's on the example, so I typed accordingly :
pt MEM write 4 0xFED1c224 0x81
But I got general protection fault instead.

Any help on this problem? BTW, I've solved the code 12 error by magmabox method, if that is ever relevant to this issue.

Once x1E succeeds, I will overclock the card to equal rate to the ones posted at x1E results table, and run some more benchmarks.

p.s. - CPU-Z says my cpu's(i945PM) chipset is ICHM7-M/U
p.s.2 - port2 on the everest screen cap says 'empty'

 

x1E instructions have been updated with the important point of disabling the adjacent port if your bios has it enabled, as it is in your case. Otherwise x1E won't work, likely causing a BSOD or some other Win7 error.

Same link has posted DMCV4 benchmark results which don't seem to be as CPU bound as RE5. dmcv4 DX9 has a 30% performance FPS improvement with x1E!!

 

Hi guys,

I've been busy running a number of tests and benchmarks in x1E mode. Exciting times
I've updated my results post here. Have a look and tell me what you think. I still need to upload some screenshots to back up all the listed results (ongoing), an Everest screenshot showing the expresscard port running at dummy "x2 1.0" , etc.. I'll have everything up soon.

 

Hi everyone, I've read most of these posts plus the 'lets build a vidock' thread but I still have a question.. I have a Lenovo W510, 4GB RAM (Model 43194BU) running XP 32bit. I really only want to use this external card for 1 or 2 monitors for work use(having 3-4 monitors total - 2 from docking station + 2 from this diy ViDock), I won't be playing any games.. just programming and systems monitoring etc..

My question is, if I purchase the PE4L-EC2C + Power adapter(12v-80w or the gateway 6831 fx 120watt power adapter listed above) what else would I need? what video card would work where I wouldn't need to mess with a ViDock Bootdisk? I have a PCI-e 8400GS, would that work? I read somewhere that NVidia cards cause a conflict with the internal W510 Nvidia card, so what equivalent(and cheap) ATI cards are there? Thanks in advance guys!

 

My bad. I confused 'disabled' to be the same state as 'empty.'

Strange... I did go ahead and disabled port#2, confirmed with everst.
But when I reboot or shut down and restart, and boot the usb disk, I still get the same general protection fault .
The disabled port#2 also reverts back into enabled state once I reboot it(I checked the 'Apply Save at Startup' box, too).
Do I need to use different save such as Del Sav, Writ Sav, etc?

 

The bios will revert the ports back to x1 mode. So boot DIY ViDock bootdisk and run DV8000.bat below to turn off port2, and set port1 as a x2 port, then chainload Win7. Likely will do the job for you, assuming the Magmabox driver doesn't alter this part of the configuration.

dv8000.bat

Code:

:: Disable pci-e port2 to allow port1 to work as x2 port
:: RCBA+3418h bit 17=1 disables port2
pt MEM write 4 0xFED1F418 0x3A0061

:: Set port1 as x2 port ("x1E" port)
:: RCBA+224h bit 1:0=10 for x2 port1
pt MEM write 4 0xFED1c224 0x81

:: Additional PCI Reallocation/Bridge configuration here

:: Chainload Win7 using bootmgr or MBR method
grub --config-file="menu.w7"
:: grub --config="menu.lst"

If you have a free 256MB window in PCI Configuration space OR can compact the space, then have a good candidate system to do a DIY ViDock. Can't say if you will or won't need to much with the DIY ViDOck bootdisk until you plug a desktop video card in and see how much configuration the bios does for you. A T400 owner reported a PNP setup using a 8400GS here.

Need Device Manager -> Resources By Connection -> Memory -> PCI Bus sections expanded under C/D/E/F0000000 to have an idea of what's going on.

 

I have connected external MSI ATI Radeon HD 5770 to my Sony Vaio VGN-FW45GJ laptop.



For system stability, I installed Desktop Cataylst 10.3 and Mobility Cataylst 10.3. My 3DMark2006 (default settings) score is 9829.

I used Antec True Power 430 Watts PS.

 

No luck... I created DV8000.bat file using above script, booted the usb, ran the batch file, but it still seems it is getting that general protection fault .
Do I need to edit other .bat files?

 

do you want x2 config?

or, be in your setting problem of x1?

in the case of x1, you need to set the memory allocation. you MUST allocate your-new-external-graphic-card enough memory, 256mB..

check it surely.

 

For code12 resource allocation problem, I've solved it using magmabox driver.
So my x1 config already works, but I am trying to get x1E config working.
Would it be possible if magmabox solution is causing this problem ?

 

first of all, delete magma(useful in x1 but incompatible with x2, this thread)
try it again, with usb boot of ...

 

Need that screenshot with your 4GB plugged in, rather than 2GB.

 

OK - try the following.

Code:

:: Setup pci-e port2 using C0000000
@echo off
set port=1
set io_win=8080
set mem_win=cfffc000

:: Relocate Nvidia 48MB from ccxx to f9xx to free up 256MB window at c0000000
echo Relocating Nvidia 48MB window from cc000000 to f9000000...
setpci -s 0:1.0 20.l=fbfff900

echo Configuring underlying bridge hosting desktop video card...
setpci -s 0:1c.%port% 1c.w=%io_win%
setpci -s 0:1c.%port% 24.l=%mem_win%

:: chainload into Win7 using bootmgr/mbr method
grub --config-file="menu.w7"
:: grub --config-file="menu.lst"

 

Any tests with 8400M and HD5670? I'm wondering about difference, as I'm looking for card for my PE4L.

 

new results

if anyone wants me to run anything let me know

big hug to everyone


btw the 8400gs scored 1450 or something equally embarrassing...

 

I've deleted magmabox, tried to set up bridge.bat myself since my ram is 2gb, but couldn't get it to work, so I reinstalled magmabox, and booted the disk, and ran DV8000.bat again just to check it.
Now it does not give general protection fault, but once the first 'pt MEM write...' command is run, nothing happens and computer kind of freezes(I see blinking text cursor, but I can't do anything else other than rebooting).

For doing bootdisk method instead of magmabox as told by chung_gun, I need some help figuring out io window, memory window, etc...I will soon post required infos written in Appendix 2 of the first post.
I am sorry for keep asking these annoying questions. I am trying my best to learn myself the basics, but it has been very difficult.

p.s. Attached pics are my laptop's PCI space with magmabox installed for reference.

p.s.2. I went ahead and ran stuffs mentioned in the appendix 2, but several of them cuased errors(ex. atiflash gave general protection fault)...

 

Try booting the DIY ViDock bootdisk and issue the following commands at the prompt step-by-step. I've also sandwiched in 'pt PCI map' to ensure pt can run on your system and 'lspci -tv -n' as a test just to ensure the PCI bus responds AND the system hasn't frozen. Commands are specific to DV8000's RCBA=0xFED1C001 AND your FD(3418h) register default being 0x3A0061 with port2 disabled as shown here.

Code:

pt PCI map
pt MEM write 4 0xFED1F418 0x3A0061
lspci -tv -n
pt MEM write 4 0xFED1c224 0x1
lspci -tv -n
grub --config-file="menu.w7"

Command above sets port1 as a x2 port but does NOT set it to be a high priority port. I recall testing that feature on my 2510P ages ago finding it would freeze, so maybe that is the problem? If so, will need to upgrade the x1E instructions.

Please advise if notice freeze and/or error after any of these commands. If it chainloads into Win7 OK, then can test with the Magmabox driver AND/or define the bridge manually.

 

I followed your instruction, but it froze(did not return c:/> prompt) starting from 'pt PCI map' command .
It looks like all 'pt...' commands didn't work. I didn't test ones starting 'lspci...' commands separately, but I remember 'lspci -tv > tree.txt' of bootdisk code at least did not freeze on prompt.

 

Hmm.. suggest try making a proper MS-DOS bootdisk (eg: Win95 DOS bootdisk) from allbootdisks.com, in case freedos is causing some problems. Then download and extract peridos.zip to it and try 'pt PCI map' to see if it works. If not, then need an alternative way to do the two 32-bit memory writes you need

pt MEM write 4 0xFED1F418 0x3A0061
pt MEM write 4 0xFED1c224 0x1

If you use Ubuntu 9.10 with the grub2 bootloader, can just use 'write_dword' to do this by adding a write_dword line to th Win7 menuitem in /boot/grub/grub.cfg. Eg: 'write_dword 0xFED1F418 0x3A0061'.

 

Hello nando4,

Making Win95 bootdisk, extracting contents in peridos, and booting to prompt somehow gave that similar freeze/no response once I chose that start with prompt option.
Making DOS6.22 with same procedure resulted freezing at 'Starting MS-DOS' message.
I've also tried using Virtual FDD with both peridos and DIYViDock bootdisk extracted into, which successfully booted to DOS, but with pt commands I got the same result again .
As a not-at-all-computer person I even felt like maybe I did something wrong when I was making bootdisks .
Does it mean I am out of luck for x1E config?

Thanks again for thorough help.

 

hi, prazisions.

in the case of Virtual FDD, you must exclude files, command.com, kernel.sys, kernel2.sys of the DIYViDock folder from making a bootible VFDD, 'cause VFDD uses the boot image in itself.

and VFDD is not compatible with window7, but in XP.

 

Thanks for advice. Will keeping them in separate folder do the job?

 

Performance: x2 1.0 versus x1E versus x1 1.0

Setup

ICH7M-equipped T2050-1.6 HP Pavilion DV2000 with HD5750. Using port3+port4 to do a x2 1.0 setup. Needed a mPCIe socket soldered onto the systemboard pads marked MINI2 to provide port3. This is easiest is solder flow the pads, position the mPCIe socket over the pads, then use a butane torch to effectively weld the two together being careful to not melt the plastic bits.

System	RAM	GPU	3dmark		RE5 dx10		dmcv4.scene4		PCIScript	Ports	OS
			06	vant.gpu	var	fixed	dx9	dx10
T2050-1.6 HP DV2000^1	3.0	HD5750@x2	6735	5579 +	38.5	18.3	60.1	55.4	dv2000	ICH7M	Win7/32
		HD5750@x1E	6694	4903 +	34.6	17.0	59.9	43.7		ICH7M
		HD5750@x1	6560	4219 +	33.2	16.2	43.8	36.7		ICH7M

^1 port3 required mPCIe socket to be soldered onto systemboard's pads marked MINI2
Bold - benchmarking showing x2's superior performance

I used dx10 since it is definitely more taxing on pci-e bandwidth than dx9. dmcv's scene4 has a lot of textures being shown so is a good one to compare performance. Faster P8600/T6600 x1E capable system dmcv4/dx10 results here cannot match x2 1.0 performance of the substantially slower T2050-1.6 cpu. The very texture heavy scene4 see x2 1.0 give 27% better performance than x1E and a whopping 51% better performance than x1 1.0 on the DV2000. x1E is 20% faster than x1 1.0 rendering scene4.

We can see that previous RE5 and 3dmark06 benchmarks are not indicative of DX10 gaming performance. Only 3dmark vantage and dmcv4.dx10 scene4 truly reflect how FPS increases with increases pci-e bandwidth.

Another example: NFS Shift is buttery smooth using x2 1.0, even when animating many cars using the T2050 CPU + HD5750@x2. x1E is a certainly more jumpy, even when using a faster T6600-2.2 CPU. x1 1.0 crawls with many cars shown simultaneously. The pci-e bandwidth is definitely the bottleneck with NFS Shift.

Summary

x1E is a good compromise but if your system can do x2 1.0 then it's definitely worth it for the substantially better performance especially in DX10.

 

Excellent stuff, congratulations on your mod! Your results also fit with my observations that while it's fine for DX9, x1E doesn't quite cut it for DX10 (3dmark Vantage, Crysis, Bad Company 2, RE5 DX10, etc). Bad Company 2 runs great with DX9 (low detail, 1280x1024, 40-70 fps) on my x1E setup but visibly struggles on DX10/11 (same settings, 20-30 fps). I'll have some DX10 figures up soon.

Unfortunately, the mPCIe riser card I need to attempt an x2 link is coming from Taiwan via Sweden. With the volcanic ash cloud mess going on in Northern Europe right now, I doubt I will receive it before May

 

I am booting in the same way you described(reboot->esc->boot order to usb).

Don't know if it is going to make any difference, but I will try making bootdisk out of MS-DOS 8 that comes along with XP just in case...