Got a tip for us?
Let us know
Become a MacRumors Supporter for $50/year with no ads, ability to filter front page stories, and private forums.
Ok Overclocking the HD2600 in Vista?
- Thread starter zepharus
- Start date
- Sort by reaction score
Hmm... wait and I will make something up in wikipedia and then link to it
hahaha yea its really easy to do that aye. and people would probably believe you. yea but im not sure tho, u may be right lol.
EFI is still a BIOS is some ways. It is an extended version, Bios v2.
Firmware and BIOS can be interchanged but aren't exactly the same.
The BIOS and EFI and firmware are all allocated address space in a computer. (This is why you never get 4GB's on 32bit systems even if you have 4GB of ram in there).
So basically if you know the "address" of the BIOS then you can read it. You can also directly access the GPU and it's memory via this method. You can even "address" the sound card directly this way, although if you don't know the correct data to put in the sound cards "memory" then you'll cause havok.
Due to this you can actually edit or read the graphics card BIOS/Firmware from within an OS.
Firmware and BIOS can be interchanged but aren't exactly the same.
The BIOS and EFI and firmware are all allocated address space in a computer. (This is why you never get 4GB's on 32bit systems even if you have 4GB of ram in there).
So basically if you know the "address" of the BIOS then you can read it. You can also directly access the GPU and it's memory via this method. You can even "address" the sound card directly this way, although if you don't know the correct data to put in the sound cards "memory" then you'll cause havok.
Due to this you can actually edit or read the graphics card BIOS/Firmware from within an OS.
because that was a hitch/limitation in the software.
wait hold up...nah yea that would have been limitations in hardware.
i was thinking about osx going to 64-bit, that would allow for the software to support...uumm...512gb RAM??
I think you are being a little too hard here since there was so much confusion about what the chip actually was. First IMO, under-clocking it does technically make it customized. Even an Apple rep. said it was a custom chip when questioned about the device ID, at least according to the 2600xt thread which I followed for a while because no one really knew.
I think this is because Tiger is 64-bit and the new Alu iMacs are fully 64bit. The older core duo white imacs were just 32bit.
If the core2duo white iMacs had this issue then it was a chipset problem.
Napa-based chipset (945) was 32-bit. So had a 4GB ceiling. Take away IO space and you are left with 3.3GB or so. Santa Rosa-based chipset (965) is 64-bit (well not really but at least not 32-bit) and therefore can address more than 4GB total (memory + IO) but only 4GB of memory.
Surely if you knew your information was ill informed, incorrect and not useful you wouldn't post it in the first place. And if everybody posted as if what they knew wasn't a fact but something that might wrong, the whole world would be full of gibbering timid people who daren't say anything.
There's some quite astonishing language in a couple of your posts.
Indeed I would hardly post something I knew was wrong and I wouldn't say anything at all if I had to be 100% sure as even I am only ever 99% sure of something.
Okay I think I understand now, On Napa chipsets, I/O is mapped onto the address space of the memory but on Santa-Rosa I/O is mapped separately ?
Hmm I thought Santa Rosa was Intels first full 64bit mobile chipset ?
Some versions of the 965 chipset appear to top out at 8MB of memory space. I don't know why Intel limited the physcal address lines to the memory but probably had something to do with lack of 4GB+ sodimms.
All I/O is mapped out to address space. However in 32bit chips this falls in the 4GB memory limitation. in 64 bit computers there is still I/O mapped in address space, but the address space is so massive that it does not interfere with memory.
However when the amount of memory starts to reach 2^64bits we will have the problem again. Not for another good few years hopefully!
I think the P965 has a 33bit address bus, however only 32bits might be used because of cost limitations (also maybe to cripple systems so people will want to buy new ones).
The mobile 965 has a 4GB hard limit though.
EDIT: There is an 8GB inherent limit so this would mean the bus is 33bits and not a 36bit.
EDIT2: The P965 (desktop chipset) has an 8GB limit, however the mobile 965 has only a 4GB limit.
I think everyone should post anything they want, be it right or wrong. However they should include a disclaimer such as
"I think X&Y"
or
"I'm not sure but at a wild guess i'd say W&Z"
If i don't know something i use a disclaimer or i go research it.
The 2600 in the iMac is definitely not a custom chip. It is just a run of the mill 2600 chip running slightly slower than capable. Underclocked, detuned or rebadged would be correct words for it.
There was initial confusion over what the chip was because people didn't actually know and they were making rumours. However as soon as someone got windows installed and used a tool to identify the chip then the ambiguity was gone.
I thought as much. Thanks TBi.
Maybe I should put the disclaimer in my signature
Why is it windows can recognise it as a 2600 XT yet the drivers wont install or do drivers just look at the device ID ?
AFAIK, laptop drivers are the responsibility of the laptop builder, so Apple. ATI drivers are mainly desktop so they may or may not work.
Maybe I should put the disclaimer in my signature
Why is it windows can recognise it as a 2600 XT yet the drivers wont install or do drivers just look at the device ID ?
Windows uses a .INF file to load the drivers. If the device ID of your chip is not in the .INF then windows will assume the drivers are not for your card. However if you edit the .INF to include the relevant device ID then the drivers will install and run fine. (ATi drivers usually needed a few more changes though)
I did this a lot with my old old dell which has a mobile nVidia chipset. Dell had a different chipset ID than the regular nVidia chips.
I also did this with a bluetooth adapter which was Mac compatible with pay-for windows drivers. I modified the built in windows bluetooth drivers to install the device as "TBi's Bluetooth Adapter". It is currently installed as this
EDIT: I actually used this to modify nVidia drivers to make my Riva TNT2 show up as a Riva TNT2 Ultra (i can't remember the exact name). I tricked a few friends with that
. You can still do it today, you can modify them to make your 8600GTS come up as a 8800GTX Ultra in windows.
Windows uses a .INF file to load the drivers. If the device ID of your chip is not in the .INF then windows will assume the drivers are not for your card. However if you edit the .INF to include the relevant device ID then the drivers will install and run fine. (ATi drivers usually needed a few more changes though)
I did this a lot with my old old dell which has a mobile nVidia chipset. Dell had a different chipset ID than the regular nVidia chips.
I also did this with a bluetooth adapter which was Mac compatible with pay-for windows drivers. I modified the built in windows bluetooth drivers to install the device as "TBi's Bluetooth Adapter". It is currently installed as this
That sound more plausible than my explanation.
You are right. ATi leave it up to Apple to test the drivers on their hardware and release accordingly.
However for the last few years ATi has released both mobile and desktop drivers in the same package so you can unlock them if you have the right know how. This is basically all apple does, they get a driver and test to make sure it works. If it does then they modify the install file to allow it to install.
That sound more plausible than my explanation.
Posted just after you did
I also did this with a bluetooth adapter which was Mac compatible with pay-for windows drivers. I modified the built in windows bluetooth drivers to install the device as "TBi's Bluetooth Adapter". It is currently installed as this
Out of interest I heard a while back that the elgato use hauppauge devices, could a similar thing be done so you can use elgato drivers on a hauppauge device or am i just talking nuts heh ?
First off i only know how to do this in windows. I know it is technically possible in OSX but i have no idea how to do it.
As for it actually working... It depends on how closely configured they are. If the Elgato is basically a rebadged hauppauge then yes of course the drivers can be modified.
However sometimes companies might use the same "processor" but different input/output circuitry. This might not be supported in the hauppauge drivers and so this would not work.