i was referring solely to stupidity, and that the 'Aeron symptom' is not restricted just to.com.
Eg. Compaq bought Digital for it's service wing. At DEC we worked as a tight-knit group. Though the staple calls were PCs, we'd take calls on anything and everything: printers, terminals, PC servers, VAX, networking kit, etc.. if you thought you could figure it out you were free to go spend time looking up stuff, asking other people and/or read and try figure it out - and if you couldn't you sent an engineer out.
You could also do sideline projects. Eg setup a website if it was useful. When they transitioned from the old VAX terminal-based call logging system to a windows based thing, a guy in our group wrote an application to present a windows front-end to the contract data stored in the VAX systems. I setup and adminstered the server (linux for scripting and network reliability. NT had problems holding multi-hour ftp sessions) that took big database dumps from the VAX systems across Europe and shared it out via Samba to that windows front-end. My first real-world experience of Unix systems admin.
I believe that's about good as it can get for a desktop customer technical support agent. consequently, people actually stayed with DEC. The group i was in, of about a hundred people, lost only a few people per year. which, as i found out later in compaq, is an amazing achievement in a call-centre. other, longer-standing tech-support groups had people working there with many/years/ of service! No expensive chairs needed to keep people in DEC.
Ie, people stayed with DEC. knowledge was acrued - it didn't just walk out after a year or less. and the knowledge was wide-ranging cause people were allowed scope and variety in the customer problems they dealt with.
Enter Compaq.... for about 6 months or so not much happened. life went on as usual. then they got around to 'integrating' us. we all had to do the Compaq induction course, where a video explained to us that Compaq was the major supplier of high-end financial transaction systems (eh, no. that's tandem - you just bought 'em), that compaq was at the forefront of RISC technology and Unix experience (no.. that's Digital, you bought us) aswell as the usual PC facts and figures.
Eventually, we were migrated from our cluttered but cosy DEC offices to the Compaq call centre. A big 5 story building with big open floors full of open-plan desks - it looked soulless to me, like a factory.
Initially, our old DEC group stayed together and we continued to do the same work we did. But we lost most of our test kit, old terminals, printers, pc's, DEC servers, etc.. couldn't have that cluttering up the desk farm. no instead we could use the compaq test kit - course you need a 'coach' with you to use them and course it's all just compaq PCs. Bit by bit things changed. next, they moved us all to different parts of the building. we lost more privileges. no longer did we have adminstrator access to our own PCs. once they tried to make everyone keep their jackets and personal belongings in a cloak-room at the beginning of each floor each morning - as if we were kids in primary school!
The final insult was when they completely integrated the (remaining) people with the original Compaq people. Compaq support is divided into Consumer (presario), Business (deskpro) and Enterprise (servers and anything high-end). Despite reassurances that we'd be integrated into the appropriate group depending on ability, they just lumped us all into various parts of the Business group - including the DEC second level guys! We were people who used to take calls on and troubleshoot problems with terminals, servers, print-servers and more !! A lot of us even had done the Compaq ACE and ASE courses - cause digital did support for Q even before they were bought out! but we were all put in the "did you reinstall windows?" group.
i'd already gone by then though...
they bought us and then slowly stripped away everything that made taking support calls at DEC a job worth being interested in. until we were either drones who apathetically took their X number of calls per day in between browsing job sites - like the compaq classic people... or we had left.
compaq's call centre in dublin has Aeron chairs at every desk. 4 floors with hundreds of desks on each floor..
it was one of the things they tried to impress the DECies with when they were integrating us into the Q. Strange... most of the DECies have left, Q has dropped Alpha..
Very much agree with you... That's why Sun have always done well, because they have always shown utmost commitment to their arch and OS - and the *nix market appreciates that.
(i don't like solaris though)
However, both SGI and HP have committed themselves to dropping their historical home-grown CPU's and moving to IA64. So cross them off your list. So that leaves Sun and IBM. (The latter with the most god-awful OS that purports to be Unix ever. Dear God let IBM hurry up with their Linux plans.
even "production super cars" will be soundly beaten by bikes that cost a fraction of what they do.
Bikes like the new Suzi GSXR1000 are extremely fast, 150 odd HP, 170Kg. Sub 10s standing start 1/4 mile times are quite common for big bikes. And even smaller 600cc bikes will hit 13/12/11s times.
I have an Aprilia RS250, even it shames mid-range ferraris and porsches for acceleration to 60 and 100mph.
cars are just slow slow slow... even the really fast ones are slow compared to bikes - and you can only ever dream about owning those.
that's not clarifying what you said. that's changing what you said.
Linux runs fine on MIPS, and has done for years. The Cobalt Cubes were MIPS based until Sun bought them, so Linux/MIPS has been at the core of a commercial product for many/years/.
is there a timeline for trying to get XFS into the stock linus tree?
as a user, i'd really love to have XFS. I use it on SGI and think it is excellent. However, I have usually have several patches to apply to my kernels, eg LVM being one, and XFS is so big and touches so much that i doubt that i will add it into my local src/patches dir.:)
So when will you submit it to Linus? Hopefully ASAP so that i can try it out ASAP.:)
The only SGI systems that run Linux right now are the low-end Intel workstations
balderdash... SGI have people working on MIPS and they have linux running on Origin2000 with *128* CPUs. Granted, the work seems primarily intended to develop linux so that it can be suitable for use on the IA64 O3000s, but you're still talking bollocks.
mainly because with IPv6 they have a clean slate to assign the addresses properly to allow for clean and dense aggregation.
also, ipv6 allows for things like having the assigning the last 64bits of your address space statically or dynamically but/without/ having it tied to the first 48 bits (or whatever) that controls how packets are routed to you. IPv6 DNS also supports this division of host id and network. So that you can renumber your network from DNS by just changing one record!
What it means is that, where currently with IPv4 if you want 2 redundant links to the internet from 2 providers you will have to either:
- get a provider independent chunk of addresses from your NIC and have both your ISPs add this (small) subnet to their BGP adverts. PI subnets are increasingly more difficult to get, cause they're running low and cause they are a huge overhead on routing at backbones, and hence discouraged - your ISP isn't even required to add your PI to their BGP adverts.
- or get provider dependent subnet from one and persuade the other ISP to advertise this chunk (not good).
- or get a PD subnet from both and dual-home all your hosts which could be a mighty pain in the arse if you have any significant number of hosts.
Instead, with IPv6, you just get 2 chunks of address space, say dead:beaf:: and f00b:a43d:: from each of your ISPs address space. You assign a unique host id to each machine, and let them figure out that their full ipv6 addr can be either dead:beaf::hostid or f00b:a43d::hostid statically, or even better, dynamically from the peers/dhcp server/ routers around them.
And v6 DNS supports this fully, you look up an A6 record and the answer consists of hostportion and a pointer to which records to look up to find out what the network portion is. You look up the network record, cat the previously found host portion and this network portion together and you have your IPv6 address.
(ie change just that one network record and you've updated network number in DNS for all your hosts, cool).
anyway, sorry i can't be more specific about the ipv6 auto-config stuff, but it is in the specs. they did think about this stuff over the last, what, 8 years or so (???) that they've been working towards ipv6.
If people are interested in playing with IPv6, well play with it at home! Eg, linux with the USAGI patches (www.linux-ipv6.org) works perfectly. Then you can get a tunnel from the 6bone (6bone.net), and after that maybe even a/64 from your tunnel provider. Ie: public IP addresses for all your machines at home - bye bye NAT!!
presumably so that it would be really easy to remember so that there could be no confusion amongst the troops as to when the Armistice came into effect.
since then SGI have booted linux on a 64 node Origin2000.
--paulj
Re:X protocol chattiness is the problem not bandwi
on
Low-Bandwidth X
·
· Score: 2
Well, we actually use Citrix+MS Terminal Server in work too. And by god does that absolutely suck.
The biggest problem is the Terminal Server side of things, it requires heaps of RAM cause multi-user is obviously just a gross hack - each user is just assigned a big block of memory, and if they all run excel, each users address space has to have the text loaded in. We have an NT box with half a GB of RAM just so that about 20 people can run word/excel/proprietary windows app.
no sharing of text between different users. ugg..
the citrix display exporting side of things works quite well (though it can very occassionally hang the display). but then, X works/really/ well - its a lot faster than citrix/TS on a LAN.
I notice you didn't bother to even try justify your claim of "[X] is an old tired technology". what specifically needs improving? Anti-aliasing extension has been written and implemented for XF4, this very article is about a new X compressor. So what do you mean with "never improved"?
from the non-technical POV, citrix+MS TS is an absolute nightmare in terms of licensing. you pay for MS the local NT workstation client, for TS and per-user licence for TS. Plus citrix per user on top of that...
uggg..
--paulj
PS: nice troll, but not good enough.
Re:X protocol chattiness is the problem not bandwi
on
Low-Bandwidth X
·
· Score: 1
congratulations! an informative post in a slashdot article about X - wow! (where are all the "X sucks" posts today?)
The problem with X over non-LAN links is indeed latency - not bandwidth. X is not a fat protocol, it's very thin actually[1], but it is extremely chatty. You could have the fattest pipe on the planet to your X client but it would still suck if the latency was bad. (eg high-throughput, high-latency satelite links).
Small applications tend to be fine. eg xterm is fine over a 33.6k modem, but the bigger the application gets, the more windows and child windows it has, the more chatting the Xserver and client have to do. Plus some apps just have to do needless fancy stuff (eg as someone posted, the gimp has an animated border for selects) which kill performance on high-latency links.
X: it doesn't suck, despite what the kiddies on slashdot might say.
[1]. data that it transfers, eg icons, is larger than it needs to be as they are raw {pixmaps,bitmaps} but server caches these.
reason i asked the question was cause i didn't what the D stood for! And i just knew someone was bound to answer.:) I didn't know it was Fabrieken though, i assumed it was just Fabriek.
en ik ben eigenlijk ook nederlands, of tenminste half-nederlands.:)
no he isn't. that's Andre's usual style that we see regularly on linux-kernel, very blunt and to the point in the fewest chars possible.
however, whether you like his atypical style or not, he definitely knows his stuff - he maintains all the linux-ide drivers, implemented most of the newer stuff in them (ATA33/66/100, serial ATA, etc..) and was invited to sit on the T13 committee that decides on future ATA standards - a committee typically composed of representatives from various vendors.
you can rest assured he didn't get that seat cause the committee were looking for a dealer.:)
When you talk about server-rooms, instead of ye old pentium box doing some printer sharing 250watts becomes small.
absolute balderdash. In your data centre power usage is/more/ of a concern than a PC at home or in a small office. You have to be careful and make sure you don't draw more power from a rack distribution box than it is rated for.
Then you have to make sure that the peak power drawn by distribution boxes attached to one of your UPS line doesn't exceed what it is rated for.
Then you have to make sure the total peak draw on your UPS is below its spec.
Then you have to make sure you have enough mains 3 phase to feed your UPS at peak output.
Then you have to make sure that your generators can supply the UPS at peak.
So if equipment such as servers were to suddenly jump in power requirement from ~700W peak to maybe 2.5kW peak, this would poke a huge hole in your power planning.
yep, alpha can use standard PC parts. ('cept soundcards which have a devious tendency to not implement a few of the higher order address lines out of cheapness).
Cheap though they are not. In fact they're damm expensive new - you'll never get a 21264 for anything less than 3k.
However, the previous generation of Alpha can still be had secondhand, even from some vendors, for a much more reasonably price. Eg you could get a 600MHz 21164A for about IEP£1k to £1.5k, or motherboard+chip for about £600.
Re:Clearing up some points
on
3Dwm Updates
·
· Score: 1
dri X specific? no it's not...
DRI is an arbitration manager for direct rendering. anyone can use it, not just X...
so? XF4 is totally different code to 3.3.6. At least try running it clocked normally before complaining. Random SIGSEGVs are a classic sign of hardware problems.
see the same thing on linux-kernel lots of:
"latest kernel x.y.z crashes on my OC'd system. It can't be the OC'ing cause it's been running fine with older kernels"
and when they do go and try it with a normally clocked processor 99/100 times it then works fine.
Slashdot Mirror
i was referring solely to stupidity, and that the 'Aeron symptom' is not restricted just to .com.
/years/ of service! No expensive chairs needed to keep people in DEC.
Eg. Compaq bought Digital for it's service wing. At DEC we worked as a tight-knit group. Though the staple calls were PCs, we'd take calls on anything and everything: printers, terminals, PC servers, VAX, networking kit, etc.. if you thought you could figure it out you were free to go spend time looking up stuff, asking other people and/or read and try figure it out - and if you couldn't you sent an engineer out.
You could also do sideline projects. Eg setup a website if it was useful. When they transitioned from the old VAX terminal-based call logging system to a windows based thing, a guy in our group wrote an application to present a windows front-end to the contract data stored in the VAX systems. I setup and adminstered the server (linux for scripting and network reliability. NT had problems holding multi-hour ftp sessions) that took big database dumps from the VAX systems across Europe and shared it out via Samba to that windows front-end. My first real-world experience of Unix systems admin.
I believe that's about good as it can get for a desktop customer technical support agent. consequently, people actually stayed with DEC. The group i was in, of about a hundred people, lost only a few people per year. which, as i found out later in compaq, is an amazing achievement in a call-centre. other, longer-standing tech-support groups had people working there with many
Ie, people stayed with DEC. knowledge was acrued - it didn't just walk out after a year or less. and the knowledge was wide-ranging cause people were allowed scope and variety in the customer problems they dealt with.
Enter Compaq.... for about 6 months or so not much happened. life went on as usual. then they got around to 'integrating' us. we all had to do the Compaq induction course, where a video explained to us that Compaq was the major supplier of high-end financial transaction systems (eh, no. that's tandem - you just bought 'em), that compaq was at the forefront of RISC technology and Unix experience (no.. that's Digital, you bought us) aswell as the usual PC facts and figures.
Eventually, we were migrated from our cluttered but cosy DEC offices to the Compaq call centre. A big 5 story building with big open floors full of open-plan desks - it looked soulless to me, like a factory.
Initially, our old DEC group stayed together and we continued to do the same work we did. But we lost most of our test kit, old terminals, printers, pc's, DEC servers, etc.. couldn't have that cluttering up the desk farm. no instead we could use the compaq test kit - course you need a 'coach' with you to use them and course it's all just compaq PCs. Bit by bit things changed. next, they moved us all to different parts of the building. we lost more privileges. no longer did we have adminstrator access to our own PCs. once they tried to make everyone keep their jackets and personal belongings in a cloak-room at the beginning of each floor each morning - as if we were kids in primary school!
The final insult was when they completely integrated the (remaining) people with the original Compaq people. Compaq support is divided into Consumer (presario), Business (deskpro) and Enterprise (servers and anything high-end). Despite reassurances that we'd be integrated into the appropriate group depending on ability, they just lumped us all into various parts of the Business group - including the DEC second level guys! We were people who used to take calls on and troubleshoot problems with terminals, servers, print-servers and more !! A lot of us even had done the Compaq ACE and ASE courses - cause digital did support for Q even before they were bought out! but we were all put in the "did you reinstall windows?" group.
i'd already gone by then though...
they bought us and then slowly stripped away everything that made taking support calls at DEC a job worth being interested in. until we were either drones who apathetically took their X number of calls per day in between browsing job sites - like the compaq classic people... or we had left.
but they had $700 chairs at each desk....
except they didn't annihilate Gaul.
Caesar beat Vercingetorix after a year-long siege and accepted Vercingetorix's surrender. (vercingetorix then went to Rome to provide 'entertainment')
conquered Gaul, yes.. annihilitated - no way.
compaq's call centre in dublin has Aeron chairs at every desk. 4 floors with hundreds of desks on each floor..
.com thing.
it was one of the things they tried to impress the DECies with when they were integrating us into the Q. Strange... most of the DECies have left, Q has dropped Alpha..
Aeron chair stupidity -> not just a
The Romans annihilated Gaul? eh... when?
go look for olvwm
Very much agree with you... That's why Sun have always done well, because they have always shown utmost commitment to their arch and OS - and the *nix market appreciates that.
(i don't like solaris though)
However, both SGI and HP have committed themselves to dropping their historical home-grown CPU's and moving to IA64. So cross them off your list. So that leaves Sun and IBM. (The latter with the most god-awful OS that purports to be Unix ever. Dear God let IBM hurry up with their Linux plans.
even "production super cars" will be soundly beaten by bikes that cost a fraction of what they do.
Bikes like the new Suzi GSXR1000 are extremely fast, 150 odd HP, 170Kg. Sub 10s standing start 1/4 mile times are quite common for big bikes. And even smaller 600cc bikes will hit 13/12/11s times.
I have an Aprilia RS250, even it shames mid-range ferraris and porsches for acceleration to 60 and 100mph.
cars are just slow slow slow... even the really fast ones are slow compared to bikes - and you can only ever dream about owning those.
Bikes: Outrageously fast for very little money.
excellent... it will benefit users and yourselves to get into the linus tree.
thanks for linux XFS!
--paulj
that's not clarifying what you said. that's changing what you said.
/years/.
Linux runs fine on MIPS, and has done for years. The Cobalt Cubes were MIPS based until Sun bought them, so Linux/MIPS has been at the core of a commercial product for many
is there a timeline for trying to get XFS into the stock linus tree?
:)
:)
as a user, i'd really love to have XFS. I use it on SGI and think it is excellent. However, I have usually have several patches to apply to my kernels, eg LVM being one, and XFS is so big and touches so much that i doubt that i will add it into my local src/patches dir.
So when will you submit it to Linus? Hopefully ASAP so that i can try it out ASAP.
--paulj
The only SGI systems that run Linux right now are the low-end Intel workstations
l oad/mips128.announce
l oad/mips64.announce
/low-end/ MIPS boxen too. Eg, handhelds, embedded boards, etc. Also runs fine on my MIPS R4k Indy...
balderdash... SGI have people working on MIPS and they have linux running on Origin2000 with *128* CPUs. Granted, the work seems primarily intended to develop linux so that it can be suitable for use on the IA64 O3000s, but you're still talking bollocks.
See:
http://oss.sgi.com/projects/LinuxScalability/down
and
http://oss.sgi.com/projects/LinuxScalability/down
And linux also runs fine on many
--paulj
mainly because with IPv6 they have a clean slate to assign the addresses properly to allow for clean and dense aggregation.
/without/ having it tied to the first 48 bits (or whatever) that controls how packets are routed to you. IPv6 DNS also supports this division of host id and network. So that you can renumber your network from DNS by just changing one record!
/64 from your tunnel provider. Ie: public IP addresses for all your machines at home - bye bye NAT!!
also, ipv6 allows for things like having the assigning the last 64bits of your address space statically or dynamically but
What it means is that, where currently with IPv4 if you want 2 redundant links to the internet from 2 providers you will have to either:
- get a provider independent chunk of addresses from your NIC and have both your ISPs add this (small) subnet to their BGP adverts. PI subnets are increasingly more difficult to get, cause they're running low and cause they are a huge overhead on routing at backbones, and hence discouraged - your ISP isn't even required to add your PI to their BGP adverts.
- or get provider dependent subnet from one and persuade the other ISP to advertise this chunk (not good).
- or get a PD subnet from both and dual-home all your hosts which could be a mighty pain in the arse if you have any significant number of hosts.
Instead, with IPv6, you just get 2 chunks of address space, say dead:beaf:: and f00b:a43d:: from each of your ISPs address space. You assign a unique host id to each machine, and let them figure out that their full ipv6 addr can be either dead:beaf::hostid or f00b:a43d::hostid statically, or even better, dynamically from the peers/dhcp server/ routers around them.
And v6 DNS supports this fully, you look up an A6 record and the answer consists of hostportion and a pointer to which records to look up to find out what the network portion is. You look up the network record, cat the previously found host portion and this network portion together and you have your IPv6 address.
(ie change just that one network record and you've updated network number in DNS for all your hosts, cool).
anyway, sorry i can't be more specific about the ipv6 auto-config stuff, but it is in the specs. they did think about this stuff over the last, what, 8 years or so (???) that they've been working towards ipv6.
If people are interested in playing with IPv6, well play with it at home! Eg, linux with the USAGI patches (www.linux-ipv6.org) works perfectly. Then you can get a tunnel from the 6bone (6bone.net), and after that maybe even a
presumably so that it would be really easy to remember so that there could be no confusion amongst the troops as to when the Armistice came into effect.
ehmmm.... SMP and NUMA are orthogonal.
SMP describes processing architecture, NUMA describes memory architecture. (both in the abstract sense).
As for memory locality, linux has supported this for quite a while. SGI did the work precisely for this reason.
--paulj
and just to correct myself before anyone else does:
/cpu/ Origin2000.
SGI booted linux on a 64
since then SGI have booted linux on a 64 node Origin2000.
--paulj
Well, we actually use Citrix+MS Terminal Server in work too. And by god does that absolutely suck.
/really/ well - its a lot faster than citrix/TS on a LAN.
The biggest problem is the Terminal Server side of things, it requires heaps of RAM cause multi-user is obviously just a gross hack - each user is just assigned a big block of memory, and if they all run excel, each users address space has to have the text loaded in. We have an NT box with half a GB of RAM just so that about 20 people can run word/excel/proprietary windows app.
no sharing of text between different users. ugg..
the citrix display exporting side of things works quite well (though it can very occassionally hang the display). but then, X works
I notice you didn't bother to even try justify your claim of "[X] is an old tired technology". what specifically needs improving? Anti-aliasing extension has been written and implemented for XF4, this very article is about a new X compressor. So what do you mean with "never improved"?
from the non-technical POV, citrix+MS TS is an absolute nightmare in terms of licensing. you pay for MS the local NT workstation client, for TS and per-user licence for TS. Plus citrix per user on top of that...
uggg..
--paulj
PS: nice troll, but not good enough.
congratulations! an informative post in a slashdot article about X - wow! (where are all the "X sucks" posts today?)
The problem with X over non-LAN links is indeed latency - not bandwidth. X is not a fat protocol, it's very thin actually[1], but it is extremely chatty. You could have the fattest pipe on the planet to your X client but it would still suck if the latency was bad. (eg high-throughput, high-latency satelite links).
Small applications tend to be fine. eg xterm is fine over a 33.6k modem, but the bigger the application gets, the more windows and child windows it has, the more chatting the Xserver and client have to do. Plus some apps just have to do needless fancy stuff (eg as someone posted, the gimp has an animated border for selects) which kill performance on high-latency links.
X: it doesn't suck, despite what the kiddies on slashdot might say.
[1]. data that it transfers, eg icons, is larger than it needs to be as they are raw {pixmaps,bitmaps} but server caches these.
reason i asked the question was cause i didn't what the D stood for! And i just knew someone was bound to answer. :) I didn't know it was Fabrieken though, i assumed it was just Fabriek.
:)
en ik ben eigenlijk ook nederlands, of tenminste half-nederlands.
thanks!
Bayerische Motor Werken or something very similar (car fanatics or german readers will correct me i'm sure) approx == Bayern Car/Automotive Factory.
Next up on the automotive acronym quiz: what does DAF stand for?
But whose interested in cars anyway? motorbikes are far faster.... and a lot cheaper.
no he isn't. that's Andre's usual style that we see regularly on linux-kernel, very blunt and to the point in the fewest chars possible.
:)
however, whether you like his atypical style or not, he definitely knows his stuff - he maintains all the linux-ide drivers, implemented most of the newer stuff in them (ATA33/66/100, serial ATA, etc..) and was invited to sit on the T13 committee that decides on future ATA standards - a committee typically composed of representatives from various vendors.
you can rest assured he didn't get that seat cause the committee were looking for a dealer.
When you talk about server-rooms, instead of ye old pentium box doing some printer sharing 250watts becomes small.
/more/ of a concern than a PC at home or in a small office. You have to be careful and make sure you don't draw more power from a rack distribution box than it is rated for.
absolute balderdash. In your data centre power usage is
Then you have to make sure that the peak power drawn by distribution boxes attached to one of your UPS line doesn't exceed what it is rated for.
Then you have to make sure the total peak draw on your UPS is below its spec.
Then you have to make sure you have enough mains 3 phase to feed your UPS at peak output.
Then you have to make sure that your generators can supply the UPS at peak.
So if equipment such as servers were to suddenly jump in power requirement from ~700W peak to maybe 2.5kW peak, this would poke a huge hole in your power planning.
yep, alpha can use standard PC parts. ('cept soundcards which have a devious tendency to not implement a few of the higher order address lines out of cheapness).
Cheap though they are not. In fact they're damm expensive new - you'll never get a 21264 for anything less than 3k.
However, the previous generation of Alpha can still be had secondhand, even from some vendors, for a much more reasonably price. Eg you could get a 600MHz 21164A for about IEP£1k to £1.5k, or motherboard+chip for about £600.
dri X specific? no it's not...
DRI is an arbitration manager for direct rendering. anyone can use it, not just X...
so? XF4 is totally different code to 3.3.6. At least try running it clocked normally before complaining. Random SIGSEGVs are a classic sign of hardware problems.
see the same thing on linux-kernel lots of:
"latest kernel x.y.z crashes on my OC'd system. It can't be the OC'ing cause it's been running fine with older kernels"
and when they do go and try it with a normally clocked processor 99/100 times it then works fine.
so just go try it...