Build times can be both memory and ALU intense. AMDs K8 core wins against Netburst on both counts.
The reason netburst doesn't scale well is because the clock multiplier required to account for the loss in performance is not a simple 10% or 25%. In my last toorcon talk I compared an 540J to a 4200+. I found that while the 540 had a 1.45x clock advantage many algorithms took over 1.6x-2x more cycles (often with bignum routines 4x or more).
So against a 2.2Ghz Athlon you'd need roughly a 4.4Ghz P4 with a heck of a cache (to avoid memory bottlenecks). Against the 2.6Ghz or 2.8Ghz cores it's even more insane. That they go from 3.2 to 3.4, 3.6 and 3.7Ghz doesn't really amount to much.
Things get worse in MP systems. Comparing 2P systems it's clear to see that for two isolated tasks performance on Opteron is a hell of a lot better. Specially with a NUMA aware OS.
My 4200+ at 2.2Ghz would routinely beat out a 3.2Ghz Prescott during builds. I imagine the 2.6Ghz cores give the 3.6Ghz Intel parts a run for their money. I know for my pair of 885s I can get build times for various projects that is insanely fast compared to my 820.
So that a new core from Intel is beating their crappy netburst core is not that surprising and not a problem for AMD [imho].
I remember a level 1 course where we had to take our computer apart, identify all the parts and put it back together (it had to work at the end). While we did wear static straps they did detach from time to time and they're not perfect anyways.
Oddly enough in that lab every semseter had a few PCI boards here and there that "stopped working". I can only imagine what that is about.
"That sort of underlines the problem with the main point though--any other hardware can go anyways, why is an overclocked CPU of particular concern?"
Because you routinely have to overvolt something to run faster.
Here's EE 101 for this topic...
The frequency is limited by the switching frequency and length of the critical path. The longer the path the more time it takes to charge the wires to get stable transitions. So the remedy is to raise the voltage (hint: think VIR triangle).
So if a circuit (or an instance of the circuit) is rated at say 1.3v @ 2.66Ghz then to run it at 4Ghz you probably have to raise the voltage (unless the limitation to 2.66Ghz was artificial).
Now that we are raising the voltage it's not hard to imagine why it could break the processor.
As for your comment between 3 vs 3.2Ghz processors there are important things you're missing. Processor production is NOT exact. Not only can a run of processors on the same wafer behave differently (hint: is yield 100%?) but between runs the actual process (chemical makeup) can change as they optimize the process.
So no, a 3.2Ghz processor even though it's probably made from the same process that makes 3Ghz shouldn't be suspect.
You'd be surprised how much "wrong" can happen before you crash. I had faulty ram a while back [well more so it had the wrong timing] and it would boot, run for a while then randomly something would segfault. Then the kernel would panic and lock up, etc.
A simple op like
MOV EAX,[EBX+13]
could excute as
MOV EAX,[EBX+14]
and not result in a significant problem.
As for the self-checks and diodes. You don't have to overheat a circuit to kill it. Over volting a transistor can denature it and you'd never notice. Just like ESD could "partially break" a circuit.
In fact if you looked at a comp lab with open computers chance are at least one IC has some form of ESD damage.
I'm not disagreeing for right now. I was responding to your comment about leaving AMDland.
While I think AMD still has the winner [specially when you go 2P or higher with the HT links] Intel is improving their cores in the Memrom/Conroe/Woodcrest parts for next year.
Now if Intel got off their collective arses and got something like HT in their processors next year it would be more neck and neck.
You should look at the MCW series next year. They'll give AMD a good run I think [*]. The basics of core from the public slides I've seen is that the new cores for Intel will be wider than deep and get good MIPS/watt.
[*] I'm saying that as Tom and not "Tom the employee". I'm basing my opinions on publicly avaiable documentation only. (e.g. Rev E6 Opterons vs. speculative Intel cores).
You could have just put a 1Ghz Duron in it. They're usually really cool (ironically enough because they're a product of a refined process so they're usually better than the pricier older cores) and super hella cheap, or were.
If you are really concerned try hunting down a 1G duron.
Unfortunately Intel isn't always smart about OEM/retail parts you can buy at a store.
My 820 Smithfield had a circular base heat sink. Naturally the cpu exploded. I bought a custom copper block heatsink and it ran fine for nearly a year [before I sold it].
But yeah, a whack of low end cpus burning up is a good sign they're being damaged by the customer.
Get an 3800+ it's DC and can clock down to 1Ghz when idle [at 1.1v].
They take very little power [make little noise]. With proper ventilation through the case you can have a quite box which can kick some power from time to time.
Who says it's not damaged already. That's one of the problems with these hacks. You could break a transistor and instead of getting a 1 in 10^-20 chance of error it's now upto 10^-9. Once in a while you'll get an error, probably not notice it yourself but something your doing could be affected.
This hack may be ok for a gaming rig, but I wouldn't do it to my workstation.
You're welcome. My bias certainly is towards smaller systems (despite my overpowered workstation hehehe). So when I wrote the projects I always tried to balance features and usability with flexibility.
Didn't you know? It allows you to break all sorts of civil law so long as the truthiness is intact.
There are "whistleblower" statuses which basically involves stuff of public good that won't otherwise come out [e.g. insecure banking, pills that are unsafe, etc].
I won't pretend to know the facts of the case. Just chiming in to say Free Kevin!
1. Get an LCD monitor. They weigh much less and are easier to lug around.
2. Grow some muscles. Carrying a 50lbs computer from a taxi or car to the table is not exactly labour intensive.
3. LAN parties are dorky anyways. Usually sausagefests and in desperate need of face-to-face interaction.:-)
When I go out to a friends place I'd rather play a console where instead of being buried behind a monitor you're able to see their face and the reaction as you head shot them in Halo 2:-)
I prove my worth as a developer [*] by what I do not what I hack together at a conference.
Also I think America [and Canada] got over, at least for us techies, certainly not for marketting purposes [arrg] the whole "gee whiz bang we're working with computers". The "Hackers" era has long since died.
That isn't to say there isn't the culture around though. Codecon, shmoocon, toorcon, defcon, etc are all around and surviving (the latter being one of the biggest).
[*] There is a difference between developer and coder that often gets missed. A coder is analogous to a contract construction worker. They know how to hammer or weld or rivet but don't ask them to design the building. A software developer is analogous to the engineer. Seeing the bigger picture, taking in requirements, designing a solution, testing [proving] procedures, etc.
Seriously. You may be rolling in the dough but for me I have to justify the cost. My laptop (my dell == personal, fujitsu == work) is meant to go with me when I do contract work. I use it to earn money and hence the purchase. If all I wanted to do was gaming I'd buy a PSP [and I did] and bring that over. It can do Wifi gaming, is much lighter and cheaper.
That and a 15lbs laptop is stupid. That's really heavy to carry around (I suggest you put that, your adapter and a few books in a knapsack and carry that around an airport for 3 hours) and would run off batteries for what, like 30 seconds?
I not saying that all laptops are for is work. I read slashdot, play the occasional flash game and chat on my laptops. If you want to shell out serious dough for it go for it. all I'm saying is personally I don't see the desire.
These are the people who skateboard around and paint army logos on their keyboards ala "Hackers".:-)
Personally I got a non-descript Titan 550 case for my 2P setup. It may be one of the fastest workstations on the planet but it sure doesn't look like it:-)
If I want a serious gaming rig it'll be a desktop where I can upgrade parts without paying a serious fortune or having to mail my rig back to Dell for upgrades.
Laptops are for work. Work involves travel. Lighter is better. Screw having a 250W laptop capable of playing Oblivion at 2560x1280 at 60fps if I can't lug it around to some random meeting half way around the world.
Oh, this is written on a Fujitsu Lifebook. A laptop that gets 7 hours of battery life, runs both linux and windows, weighs about 4 lbs and fits easily in my knapsack. I like my Insprion Dell laptop [630m] but compared to the lifebook it's a monster. Weighs nearly 8lbs and while it fits in my bag as well it's a bit more cumbersome to lug around.
Besides, it's an open format. Suppose Openoffice lacks screen reader support. Someone can and may just write a compatible application for the task reading ODF as it's... open...
The argument is a non-starter for that reason alone.
You know for a fact the average "sales" guy would rape your unborn child to make a higher profit next year.
What you will see happening is groups will form and break alliances to screw around with the competition. The net result will be the customers will get raped, lose the ability to decide for them fucking selves what they want and the QoS will suffer to make "administrative updates".
How about MCI charges Verizon zero for IP traffic and in retrn Verizon charges MCI nothing as well.
If you're big enough to be a peering partner chances are you'll have enough use to warrant a bidirectional agreement.
The problem becomes when you cross your ISP borders.
I setup shop with MCI and pay 1 cent per GB. But then a customer on Verizon wants to talk to me and they throttle my traffic until I pay Verizon more money.
The problem is this is petty, greedy and doesn't scale. You get a M.A.D. situation and frankly it'd be funny to watch.
It's bad enough they're effectively double dipping making both sides of the communication pay, now they want to tax you based on where your customers are?
That'd be like paying more on your telephone bill because you got a long distance INCOMING call.
Slashdot Mirror
Build times can be both memory and ALU intense. AMDs K8 core wins against Netburst on both counts.
The reason netburst doesn't scale well is because the clock multiplier required to account for the loss in performance is not a simple 10% or 25%. In my last toorcon talk I compared an 540J to a 4200+. I found that while the 540 had a 1.45x clock advantage many algorithms took over 1.6x-2x more cycles (often with bignum routines 4x or more).
So against a 2.2Ghz Athlon you'd need roughly a 4.4Ghz P4 with a heck of a cache (to avoid memory bottlenecks). Against the 2.6Ghz or 2.8Ghz cores it's even more insane. That they go from 3.2 to 3.4, 3.6 and 3.7Ghz doesn't really amount to much.
Things get worse in MP systems. Comparing 2P systems it's clear to see that for two isolated tasks performance on Opteron is a hell of a lot better. Specially with a NUMA aware OS.
Tom
My 4200+ at 2.2Ghz would routinely beat out a 3.2Ghz Prescott during builds. I imagine the 2.6Ghz cores give the 3.6Ghz Intel parts a run for their money. I know for my pair of 885s I can get build times for various projects that is insanely fast compared to my 820.
So that a new core from Intel is beating their crappy netburst core is not that surprising and not a problem for AMD [imho].
Tom
I remember a level 1 course where we had to take our computer apart, identify all the parts and put it back together (it had to work at the end). While we did wear static straps they did detach from time to time and they're not perfect anyways.
Oddly enough in that lab every semseter had a few PCI boards here and there that "stopped working". I can only imagine what that is about.
Tom
"That sort of underlines the problem with the main point though--any other hardware can go anyways, why is an overclocked CPU of particular concern?"
Because you routinely have to overvolt something to run faster.
Here's EE 101 for this topic...
The frequency is limited by the switching frequency and length of the critical path. The longer the path the more time it takes to charge the wires to get stable transitions. So the remedy is to raise the voltage (hint: think VIR triangle).
So if a circuit (or an instance of the circuit) is rated at say 1.3v @ 2.66Ghz then to run it at 4Ghz you probably have to raise the voltage (unless the limitation to 2.66Ghz was artificial).
Now that we are raising the voltage it's not hard to imagine why it could break the processor.
As for your comment between 3 vs 3.2Ghz processors there are important things you're missing. Processor production is NOT exact. Not only can a run of processors on the same wafer behave differently (hint: is yield 100%?) but between runs the actual process (chemical makeup) can change as they optimize the process.
So no, a 3.2Ghz processor even though it's probably made from the same process that makes 3Ghz shouldn't be suspect.
Tom
You laugh but I do that. My 2P box can heat up the room in a good 30 mins or so on a chilly day.
Of course then in the summer it sucks.
Tom
You'd be surprised how much "wrong" can happen before you crash. I had faulty ram a while back [well more so it had the wrong timing] and it would boot, run for a while then randomly something would segfault. Then the kernel would panic and lock up, etc.
A simple op like
MOV EAX,[EBX+13]
could excute as
MOV EAX,[EBX+14]
and not result in a significant problem.
As for the self-checks and diodes. You don't have to overheat a circuit to kill it. Over volting a transistor can denature it and you'd never notice. Just like ESD could "partially break" a circuit.
In fact if you looked at a comp lab with open computers chance are at least one IC has some form of ESD damage.
Tom
I'm not disagreeing for right now. I was responding to your comment about leaving AMDland.
While I think AMD still has the winner [specially when you go 2P or higher with the HT links] Intel is improving their cores in the Memrom/Conroe/Woodcrest parts for next year.
Now if Intel got off their collective arses and got something like HT in their processors next year it would be more neck and neck.
Tom
You should look at the MCW series next year. They'll give AMD a good run I think [*]. The basics of core from the public slides I've seen is that the new cores for Intel will be wider than deep and get good MIPS/watt.
[*] I'm saying that as Tom and not "Tom the employee". I'm basing my opinions on publicly avaiable documentation only. (e.g. Rev E6 Opterons vs. speculative Intel cores).
Tom
Factor in cooling costs for wherever you put that damn thing.
You can't deduct points for being "elite" though.
Tom
You could have just put a 1Ghz Duron in it. They're usually really cool (ironically enough because they're a product of a refined process so they're usually better than the pricier older cores) and super hella cheap, or were.
If you are really concerned try hunting down a 1G duron.
Tom
Unfortunately Intel isn't always smart about OEM/retail parts you can buy at a store.
My 820 Smithfield had a circular base heat sink. Naturally the cpu exploded. I bought a custom copper block heatsink and it ran fine for nearly a year [before I sold it].
But yeah, a whack of low end cpus burning up is a good sign they're being damaged by the customer.
Tom
Get an 3800+ it's DC and can clock down to 1Ghz when idle [at 1.1v].
They take very little power [make little noise]. With proper ventilation through the case you can have a quite box which can kick some power from time to time.
Tom
Hehehe yeah. Stupid tampa sun getting to me.
Thanks for the correction.
Tom
Who says it's not damaged already. That's one of the problems with these hacks. You could break a transistor and instead of getting a 1 in 10^-20 chance of error it's now upto 10^-9. Once in a while you'll get an error, probably not notice it yourself but something your doing could be affected.
This hack may be ok for a gaming rig, but I wouldn't do it to my workstation.
Tom
In the storage world it matters little how densely you can STORE things. It matters how densely you can READ things.
If you stacked the platters you'd get a lot of density but you can't read it because the arm won't fit between two touching plates.
Tom
You're welcome. My bias certainly is towards smaller systems (despite my overpowered workstation hehehe). So when I wrote the projects I always tried to balance features and usability with flexibility.
BTW why the AC reply?
Tom
Didn't you know? It allows you to break all sorts of civil law so long as the truthiness is intact.
There are "whistleblower" statuses which basically involves stuff of public good that won't otherwise come out [e.g. insecure banking, pills that are unsafe, etc].
I won't pretend to know the facts of the case. Just chiming in to say Free Kevin!
Tom
1. Get an LCD monitor. They weigh much less and are easier to lug around.
:-)
:-)
2. Grow some muscles. Carrying a 50lbs computer from a taxi or car to the table is not exactly labour intensive.
3. LAN parties are dorky anyways. Usually sausagefests and in desperate need of face-to-face interaction.
When I go out to a friends place I'd rather play a console where instead of being buried behind a monitor you're able to see their face and the reaction as you head shot them in Halo 2
Tom
I prove my worth as a developer [*] by what I do not what I hack together at a conference.
Also I think America [and Canada] got over, at least for us techies, certainly not for marketting purposes [arrg] the whole "gee whiz bang we're working with computers". The "Hackers" era has long since died.
That isn't to say there isn't the culture around though. Codecon, shmoocon, toorcon, defcon, etc are all around and surviving (the latter being one of the biggest).
[*] There is a difference between developer and coder that often gets missed. A coder is analogous to a contract construction worker. They know how to hammer or weld or rivet but don't ask them to design the building. A software developer is analogous to the engineer. Seeing the bigger picture, taking in requirements, designing a solution, testing [proving] procedures, etc.
Tom
What you want is a PSP. :-)
Seriously. You may be rolling in the dough but for me I have to justify the cost. My laptop (my dell == personal, fujitsu == work) is meant to go with me when I do contract work. I use it to earn money and hence the purchase. If all I wanted to do was gaming I'd buy a PSP [and I did] and bring that over. It can do Wifi gaming, is much lighter and cheaper.
That and a 15lbs laptop is stupid. That's really heavy to carry around (I suggest you put that, your adapter and a few books in a knapsack and carry that around an airport for 3 hours) and would run off batteries for what, like 30 seconds?
I not saying that all laptops are for is work. I read slashdot, play the occasional flash game and chat on my laptops. If you want to shell out serious dough for it go for it. all I'm saying is personally I don't see the desire.
Tom
These are the people who skateboard around and paint army logos on their keyboards ala "Hackers". :-)
:-)
Personally I got a non-descript Titan 550 case for my 2P setup. It may be one of the fastest workstations on the planet but it sure doesn't look like it
Tom
And we care because?
If I want a serious gaming rig it'll be a desktop where I can upgrade parts without paying a serious fortune or having to mail my rig back to Dell for upgrades.
Laptops are for work. Work involves travel. Lighter is better. Screw having a 250W laptop capable of playing Oblivion at 2560x1280 at 60fps if I can't lug it around to some random meeting half way around the world.
Oh, this is written on a Fujitsu Lifebook. A laptop that gets 7 hours of battery life, runs both linux and windows, weighs about 4 lbs and fits easily in my knapsack. I like my Insprion Dell laptop [630m] but compared to the lifebook it's a monster. Weighs nearly 8lbs and while it fits in my bag as well it's a bit more cumbersome to lug around.
Tom
Yes, "confusion".
... open...
This is pure rhetoric.
Besides, it's an open format. Suppose Openoffice lacks screen reader support. Someone can and may just write a compatible application for the task reading ODF as it's
The argument is a non-starter for that reason alone.
Tom
HEhehehe your optimism is both ignorant and cute.
You know for a fact the average "sales" guy would rape your unborn child to make a higher profit next year.
What you will see happening is groups will form and break alliances to screw around with the competition. The net result will be the customers will get raped, lose the ability to decide for them fucking selves what they want and the QoS will suffer to make "administrative updates".
How about MCI charges Verizon zero for IP traffic and in retrn Verizon charges MCI nothing as well.
If you're big enough to be a peering partner chances are you'll have enough use to warrant a bidirectional agreement.
Tom
The problem becomes when you cross your ISP borders.
I setup shop with MCI and pay 1 cent per GB. But then a customer on Verizon wants to talk to me and they throttle my traffic until I pay Verizon more money.
The problem is this is petty, greedy and doesn't scale. You get a M.A.D. situation and frankly it'd be funny to watch.
It's bad enough they're effectively double dipping making both sides of the communication pay, now they want to tax you based on where your customers are?
That'd be like paying more on your telephone bill because you got a long distance INCOMING call.
Tom