On AMD and Intel
I read some of the heated discussions in the previous thread. There are criticisms of Hector Ruiz and his generals. However, if one takes a forward looking view, AMD is poised to make a kill.
Today's competitive landscape in PC and server computing is centered around one standard: AMD64. Soon, we will see other AMD standards, such as HyperTransport, being adopted across the industry. AMD firmly commands the high ground in enterprise computing arena. Its technologies are server based, then adapted for mobile and desktop use. Intel, on the other hand, is predominantly a desktop CPU company. Opterons are built for glueless 8-P computing from day one. A K10 quad will make 32-way affordable to the masses. A 8P Opteron system has many non-coherent HT links for throughput. The highest performing Intel platform today is a 2-P 8-way Clovertown configuration which is constrained in I/O capabilities.
The delay of K10 is unfortunate and AMD has suffered setbacks financially. However, that will be soon behind AMD. Once K10 is out, there is hardly any reason for someone to stick to the ancient FSB based architecture. K10 will smash whatever Intel has in all categories. At the high end, K10 will scale linearly to 32 way, which will be 4x of Intel (2P 8 way).
posted by Sharikou, Ph. D. @ 8/22/2007 04:02:00 PM
82 Comments:
Sharikou,
That was the most sensible well rounded post by you so far. Hopefully things will get better quickly.
is that really you, sensei? no more multiple personality (such as oneexpert, penix and sometimes just plain anonymous) complex. or did you have thom check your post?
So why ignore Henri's departure douchbag?
Hahahhaha henri is the smart one he realized there is more profit in selling hotcakes as i predicted......
Intel is accepting apps for janitors i am sure henri and hector will fill the positions well they cant possibly mess moping floors up...........or can they.......
AMD when being just good enough will suffice
Sharikou, you still checking the news?? it's over, baghdad-bob is gone. It's not too late for you to quit too. Care to comment? don't cry now
Tigerton coming September 2nd, 8 days before Barcelona launch. Up to 32P for a total of 128 cores witgh each running up to 2.93Ghz.
AMD is stuck at 8P, or 32 cores at 2Ghz.
Certainly K10 will strengthen AMD's server side I agree but I am not sure yet whether it will be strong enough on the general desktop side due to the difference in speeds of the initial K10 and the likely speeds of Yorkfield and Wolfdale.
Of course it definitely will be closer, hopefully close enough either way to continue the price war to some extent.
I agree completely with Sharikou. Once K10 is out, it will smash all of Intel's FSB crippled hardware.
abinstein already pointed out that Intel's current generation of chips is severely limited by FSB. It needs to be recognized that Intel will be staying with the under performing FSB model until 2009. This leaves the entire playing field wide open for AMD.
I agree with Sharikou and whatever he says.
"AMD is aiming for its Phenom to be just that; a marvel of performance and unit sales, representing a turn-around for the company."
"Anand benches Wolfdale against Conroe.
http://www.anandtech.com/cpuchipse [...] i=3069&p=1
With all the hype I'm surprised by the results of only a 5% increase in performance. Overclocking looks promising though. They say its also not the final stepping. Wasn't too impressed with the power consumption figures. They are good but not great.
I suppose at the end of the day they are good results for what is essentially just a die shrink. Too much hype was the problem.
On another note looks promising for AMD. If Phenom can at least match those figures AMD might stay in buisness after all."
Great news... for AMD ;)
Ahhh yes the idiot fanbois coming over to justify this.
Are you guys retards........you really must be.....
For example this idiot....
rathor says: I agree with Sharikou and whatever he says.
"AMD is aiming for its Phenom to be just that; a marvel of performance and unit sales, representing a turn-around for the company."
He agrees with sharikou that phenom is going to be a marvel of performance and will turn around the company. THIS THING IS ON SLIDESHOWS THERE IS NO PROOF OF ITS PERFORMANCE ANYWHERE YET THEY ARE BOTH IN AGREEMENT!!!! Its the retard leading the blind retard over here...
And another one Penix says: Once K10 is out, it will smash all of Intel's FSB crippled hardware.
Yes i am expecting smashing performance from a 2.0Ghz part...you are truly a retard.....
And to further legitimize his stupidity......rathor says: I suppose at the end of the day they are good results for what is essentially just a die shrink. Too much hype was the problem.
This coming from a dumbass who is more than content with a 2.0Ghz barcelona hyped to be 40% faster than the fastest clowertown he is not impressed with a 5% increase with a die shrink which intel has claimed will be the case all along........
On further thought arent you the retard that traded his overclocked C2D for an X2........never mind i guess i have my answer about you being retarded.......
So lets summarize this
We have no details about barcelona this close to launch....
No idea how much it overclocks....
No idea how it performs........
No idea the power it consumes......
And yet its touted as the second coming of Christ........
Well here is a dose of reality for you delusional hypocrites.....
THE FREAKING CEO RESIGNED RITE BEFORE THE RELEASE OF A PROCESSOR THAT WE HAVE NO DETAILS FOR!!!!!
What does that mean?????? It means my quote of barcelona being the brown turd edition mite actually be true.
Final word of stupidity comes form our very retarded friend rathor again.......On another note looks promising for AMD. If Phenom can at least match those figures AMD might stay in buisness after all."
Great news... for AMD ;)
How could that be good news you ass wipes. You have no clue how barcelona performs and the freaking CEO quit rite before launch if this doesnt cast a doubt about barcelona's performance than you mite be the dumbest person ever to grace this blog minus oneexpert, sharikou and (insert any AMD fanboi name here) ........
@13ringinheat = monument of stupidity
Your life is a monument to stupidity... You are a primordial fragment from the original big bang of stupidity.
Anti Intel guy: monument of stupidity
Your life is a monument to stupidity... You are a primordial fragment from the original big bang of stupidity.
Hhahahah so am i the monument of stupidity or is my life one or am i a fragment from the original big bang of stupidity......why dont you think about it, make up your mind and let me know cuz i realize retards can be slow which also explains why you are an AMD fanboi..........
You calling me anything wont change anything end of the day
barcelona is a turd....
AMD CEO knows this and that is why he resigned....
The only thing AMD is good at is slideshows........
AMD is in the brink of collapse.......
K8 gets its ass handed to it by Core 2 duo.......
2900 XT gets its ass handed to it by nvidia.......
Thank you now continue with that important post about what part of stupid i am i will wait for your response once your brain cells comprehend the information you are trying to spew..........
i dont care about sever processors, where are the consumer models !
The Anandtech results are false. No SSE4 optimisations. With those optimizations there is 10% increase in IPC.
http://www.xbitlabs.com/news/cpu/display/20070817173238.html
What does AMD have? Barcelona Brown Turd Edition stuck at 2Ghz. They'll be lucky to hit 2.8Ghz by the time Intel has a 4GHz Penryn thanks to the revolutionary high-k 45nm process. Nehalem is also set to debut at 4Ghz for minimum. This process allows for a tenfold reduction in leakage. This is unique to Intel; AMD's 45nm comes late next year (One year behind Intel) and even then it's not high-k! When Intel said they're years ahead of the competition with the 45nm high-k process they weren't kidding.
Oneexpert keeps lying about AMD performance. In real performance Core 2 E6550 frags 4x4 3Ghz: http://www.xbitlabs.com/images/cpu/core2extreme-qx
AMD BK Q2'08.
at least Intel is not afraid of releasing benchies for A0 Silicon (do you understand what that is AMD turdbois?). One cannot find any real benchies for Barcelona, but hey you can see a few powerpoints with simulated numbers, RFLMAO
Barcelona: too little, too late, too hot, too slow, too big to yield, pathetic. AMD BK Q2 08
Once again Sharikou is only looking at the present for Intel, while looking at the future for AMD.
What a fucking moron.
Hey Sharidouche, you need to look at the future for Intel too.
Of course if you were to do that you would look like some of your fanboi morons around here. Oh wait! You already do!
He's an idiot...He completely ignores the fact that their head of marketing has resigned and will be leaving two days before Barcelona launches. (Paper launches I have no doubt)
I fully expect to see more people bailing soon. Hector's time is short I think.
@13riningheat - Henri isn't the CEO, just thought I'd point that out....
Evil_Merlin said...
Once again Sharikou is only looking at the present for Intel, while looking at the future for AMD.
Scroll up and read my post. Intel is sticking with their crippled FSB model until 2009. This is not new information, but it is obvious that you want to overlook this important fact.
The thing AMD is counting on is that they can sell QUAD to current DUAL owners for a seamless upgrade.
Most shops will not remove 1000s of servers and replace them which is why it took AMD 2 years to really have a presence in server.
Blizzard, MS, Google, et al will definitely buy tons since they already have 10s of 1000s of HP DL585\385 and Sun X-Series.
That is also why they staggered the release of Budapest; so they coul dsupply the 2P\4P market first.
I wouldn't be surprised if Phenom FX gets pushed forward since it has been shown and played with by reporters at 3GHz.
Of course the desktop models don't have to be as "guaranteed" as server so the higher clocks will work for X4 and X2.
It's been said that X2 can get to 3.4GHz with the B2 rev at 1.4V.
it's unfortunate that AMD is up against a company liek Intel as they should be at 40% share by now, which would have enabled them to push even X2 forward.
Alas, such is not the case.
BTW, Yes I know Henri left. He's been globe-hopping for five years and did the job he was hired for. More OEMs than ever are adopting AMD and their supply chain is growing.
I would even bet that EVERY last 2GHz Barcelona is already sold.
Um, you do realize that the quad-core Intel Xeons are drop in replacements too?
Christian Howell
I would even bet that EVERY last 2GHz Barcelona is already sold.
Who in their right mind would buy the first 2.0 GHz Barcys when much faster ones are only a couple months away (supposedly)?
No, I fear that the Barcy launch is going to be a flop. Yeah the performance-per-watt might be impressive, but everyone's going to wait for the faster ones because raw performance counts too.
Ignore Penix - he's stuck on FSB. If FSB was such a bottleneck (which pretty much every site out there except these douchebags agrees its not as bad as AMD claims it is)then I'm sure they'd be working to change it much faster...as it is Nehalam (sp?) is set to include CSI on the first server spin, which OMG, is *not* 2009, but 2008...
Remember penix is just another sharidouche ID...
Spaztic Pizza said...
Ignore Penix - he's stuck on FSB. If FSB was such a bottleneck...
FSB is a severe bottleneck, and abinstein did an analysis which proved it. Stop trying to pretend the problem isn't there.
I have no faith in Intel's upcoming QuickPath Interconnect. Rather than adopt the proven HyperTransport, Intel will introduce their own option. I predict this will be another Intel failure of equal or greater magnitude than USB.
penix
"I would even bet that EVERY last 2GHz Barcelona is already sold."
You mean both of them?
"FSB is a severe bottleneck, and abinstein did an analysis which proved it"
I did a little analysis also and found out that even 2P K8 is memory bottlenecked: K8 from 2 to 3GHz, Core2 from 2 to 3GHz. AMD is massively bottlenecked in 433, 459 and 470. Not much better than the masively inferior FSB of Intel, isn't it?
For the record, that is SpecFP_rate, the same benchmark Abinstein used for its scaling analysis.
"I predict this will be another Intel failure of equal or greater magnitude than USB."
I must have missed something, what is wrong with USB? It is the standard connection to use for external devices.
I predict this will be another Intel failure of equal or greater magnitude than USB.
If USB is a failure, what is a success? Firewire???
Moron!
One of the generals seems to have jumped ship.
USB failed? Woah.
No wonder AMD fanboi's have such a strange way of looking at the world.
This comment has been removed by the author.
In reply to ho ho, core2dude and evil_merlin in regards to USB:
USB is a perfect example of how Intel holds back the industry by refusing to adopt a proven and available technology, instead pushing their inferior alternative. FireWire was already developed before USB was even conceived as an option. Intel considered using FireWire at one point. Rather than license FireWire, Intel decided to create an inferior alternative and force this on the industry. Even USB 2.0 is clearly inferior to the original FireWire 400.
USB is another shining example of Intel's failure to innovate and insistence on relying on monopolistic tactics to force inferior technology on the industry.
Once again, the fanboi leaves out the little details.
1.) Licencing firewire cost money.
2.) Longer cables in USB/USB 2.0 than firewire.
3.) More devices with USB over firewire
4.) Consistent data transfer speeds with USB over that Firewire offers which typically drops off by up to 50% under heavy loads
5.) Firewire devices are typically more expensive than the like USB devices due to the complexity in controllers.
Now what were you saying asshat?
So mind telling us why someone would connect a keyboard or mouse via Firewire?
Intel Garage sale begins :-
Intel Core 2 Duo Processor: E6600 $136, E6700 $196 at Micro Center
Evil_Merlin said...
1.) Licencing firewire cost money.
It's a good thing that developing USB from scratch didn't cost a dime.
Evil_Merlin said...
2.) Longer cables in USB/USB 2.0 than firewire.
USB: 5m; FireWire: 4.5m
Half a meter, whoop dee doo! And you neglected that both can be extended significantly using repeaters. They are essentially the same and this argument irrelevant.
Evil_Merlin said...
3.) More devices with USB over firewire
USB: 127; FireWire: 63
This is a problem. I admit that I would be disappointed if I couldn't connect 127 keyboards to my computer at once.
Evil_Merlin said...
4.) Consistent data transfer speeds with USB over that Firewire offers which typically drops off by up to 50% under heavy loads
You got it backwards you fucking tool. USB performance drops under heavy load because it uses the host as the controller. FireWire has a dedicated controller which is independant of the host. This is the equivalent of ATA vs SCSI. No contest.
Evil_Merlin said...
5.) Firewire devices are typically more expensive than the like USB devices due to the complexity in controllers.
Why buy an Opteron when you could buy a Celeron? Because you get what you pay for.
Evil_Merlin said...
So mind telling us why someone would connect a keyboard or mouse via Firewire?
Why would you even ask such a stupid question? For the exact same reason you would connect a keyboard or mouse via USB. Did you forget that USB was the inferior alternative to FireWire? They do the same damn thing, except FireWire does it better.
evil_merlin, you are a fool. You should have left this one alone and tried to dodge the facts like you always do. Now you have three options:
1. Bravely admit defeat like a man.
2. Silently admit defeat like a coward.
3. Continue to pointlessly argue like a clown when it is blazingly clear you have been defeated.
Suck it bitch:
USB 2.0: http://www.ixbt.com/storage/usb-to-fw/w2kusb2.png
Firewire 400: http://www.ixbt.com/storage/usb-to-fw/w2k-fw.png
Awww... now you look like a total ass.
Wait... you've looked like one since you started posting here.
You still lose!
Oh and that doesn't mention, typically Firewire consumes more host CPU for file transfers at equivalent speeds vs. USB 2.0
You were saying fuckwit?
Penix completely ripped by evil_merlin
well done
HAHAHAHAH owned!!!
This comment has been removed by the author.
This comment has been removed by the author.
evil melin,
the links you provided shows that firewire is about 3X faster than USB.
Whats your point?
How much does a firewire port cost on a $10 mouse? How much does it cost to add a firewire port to motherboard? How much does firewire HUB cost, if there exists any? What benefits would I have by having firewire mouse instead of USB one?
Short comparison is that firewire is designed for speed, USB for simplicity and low cost. Lately USB has been catching up with firewire speeds, though eSata will likely be the replacement for firewire instead of USB.
Name a mobo with 6 firewire ports.
Argument over. It would cost stupid money.
Penryn dissapoints....Yet it's still better than Barcelona. A0 silicon.....averages made on games. Yeah cuz games are 100% reliant on the cpu.
Um never once did I say USB was faster than Firewire...
By your logic, Intel CPU's are superior to AMD's CPU's because they are faster.
Game, Point, Match.
Have a nice day!
In response to evil_merlin:
You have chosen option 3: Continue to pointlessly argue like a clown when it is blazingly clear you have been defeated.
In your rebuttal you make 2 claims:
1. FireWire performance decreases under load.
2. FireWire uses more CPU for file transfers at equivalent speeds.
Both these claims are blatantly wrong and can be proven so using the same benchmarks you provided as evidence. You have humiliated yourself and those foolish enough to back you. As a gesture of good will, I will allow you to a chance to correct your flawed analysis and retract these 2 claims. If you do not, I will be forced to point out your blatant error and humiliate you once again.
evil_merlin, you would be wise to consult ho ho for advice in this matter since he is clearly more intelligent than you.
Ho Ho said...
How much does a firewire port cost on a $10 mouse? How much does it cost to add a firewire port to motherboard? How much does firewire HUB cost, if there exists any? What benefits would I have by having firewire mouse instead of USB one?
This is a loaded question because we know that costs drops as volume increases. Because Intel snuffed FireWire out of the gate, we will never know what the cost would have been today if it were widely adopted. Consider the cost of flash memory, which has dropped in price by magnitudes due to wide adoption. Originally costing dollars per MB, it has now been reduced to a fraction of a cent per MB. FireWire chips used for integration cost approximately $1 to $2. If it had been widely adopted this cost would have been significantly cheaper due to competition between chip manufacturers.
Ho Ho said...
Short comparison is that firewire is designed for speed, USB for simplicity and low cost. Lately USB has been catching up with firewire speeds, though eSata will likely be the replacement for firewire instead of USB.
FireWire has always been aimed at performance while USB was designed to be cheap at the cost of performance. USB will never be able to catch up with FireWire due to it architecture behind it. The USB 2.0 spec states for speeds up to 480Mbps, yet it can't come close to FireWire 400's 400Mbps. FireWire 800 is nearly twice as fast as the original 400. There are no announced plans for USB 3.0.
eSATA is already beginning to become a viable alternative to USB/FireWire for external drives. It still has problems with signal degradation, but if they can be rectified it will most likely be provided side by side with USB or FireWire on external drives. eSATA will never replace USB or FireWire on anything except for external drives, simply because that's all it can do. It is not a generic connection like USB or FireWire.
I'll say it ONE MORE TIME FOR THE FUCKING RETARDED KIDDIES AMONGST US.
USB was successful because Firewire charged a license fee! It was very expensive initally too, per CONNECTOR. Funny, if Intel was trying to snuff out Firewire why did the following happen?
1.) Apple implmeneted USB WELL before they moved over to Intel.
2.) Intel's motherboards of the era were fitted with Firewire.
Neither of those sound like Intel was attempting to kill off Firewire.
lets do a little bit of history huh since you fucking braindead AMD fanbois seem to spin your own history because stuff like the facts is way too much for you to handle.
USB was originally seen as a complement to FireWire which was designed as a high-speed serial bus which could efficiently interconnect peripherals such as hard disks, audio interfaces, and video equipment. USB originally operated at a far lower data rate and used much simpler hardware, and was suitable for small peripherals such as keyboards and mice.
Intel dumped a TONNE of money into 1394a-2000. For fucks sake Intel did nearly all of the design work.
Then Apple as usual, decided, hell lets charge for it! Steve Jobs was convinced that Apple should ask for $1 per port for the patents that became part of the standard.
To put it simply, for the simpleton AMD Fanbois who fail to see the truth, USB was designed for simplicity and low cost, while FireWire was designed for high performance. Then Apple fucked it up by trying to charge a buck a port, and when Sony and the other members of the Firewire team decided they wanted the cash too, Apple was fucked and realized that they would be paying out a small fortune. So it was quickly dropped and a special Licensing Authority was started.
By then it was too late and USB was out and was found to suit the needs of most people.
Do you still want another beating fuckwit?
You'll never learn will ya?
This comment has been removed by the author.
In response to evil_merlin:
Please stop acting like a clown. Your attempt to ignore my last post completely will not go unchecked. You made two claims which are clearly false. Correct them, or admit you do not have the intelligence required to do so, and I will correct them for you.
You whine like a little bitch when whipped.
Grow a set of balls pussy.
In response to evil_merlin:
Your refusal to admit your failure is a firm mark of shameful cowardice. Correct your claims or I will assume you are too incompetent to do so.
oneexpert
AMD commits to build NY FAB of the future.
Indeed. The NY FAB will be the most advanced manufacturing facility for micro-processors in the world thanks to AMD's revolutionary APM technologies.
Intel will suffer terribly when Barcelona launches. Only AMD has a true quad core. Intel has no quad core, they have dual dual-core CPUs.
PENIX said...
In response to evil_merlin:
Your refusal to admit your failure is a firm mark of shameful cowardice. Correct your claims or I will assume you are too incompetent to do so.
Hey idiot, your first claim was that USB was a failure. Suck it up, and spread your legs, for you are owned!
core2dude said...
...your first claim was that USB was a failure...
Failure was the wrong word of choice on the matter because it is too vague and can be interpreted incorrectly. What I was trying to convey is that HyperTransport is a readily available and proven option, but Intel is developing QuickPath anyway. This is reminiscent of the early FireWire vs USB battle. The original FireWire is still a superior performing technology, which continues to best USB after more than a decade. Intel successfully pushed USB adoption and has displaced FireWire as a dominant technology. This is a success for Intel, but unfortunate because the industry suffers.
Your interference only suggests that evil_merlin can't handle this by himself. evil_merlin is a big boy now. He doesn't need his fellow fanboys to jump in and try to save him.
penix
"Because Intel snuffed FireWire out of the gate, we will never know what the cost would have been today if it were widely adopted"
Pretty much every single motherboard has at least two firewire ports. If it would be good enough to use for simple external devices as mouse, keyboard, printer, flash drive etc there would exist such items. Why there aren't?
"have been today if it were widely adopted. Consider the cost of flash memory, which has dropped in price by magnitudes due to wide adoption."
Did price drop because of adoption or it was adopted because of the price? CPUs are also widely adopted but their manufacturing costs don't decrease because of that. Advances in production lower costs.
"USB will never be able to catch up with FireWire due to it architecture behind it."
Would it need to? I can see zero reasons to why would keyboard or mouse need more bandwidth than USB 1.1 provides.
As I said, eSata will likely replace USB and wirewire as the future technology for connecting external hard drives and other devices needing massive bandwidth. It has peak throughput the same as SATA2 that is several times higher than firewire 800 could ever provide.
"It is not a generic connection like USB or FireWire"
Technically, it is. Around today next year I'll have the first GPS connected over eSata to a PC. Currently a plain old serial port or USB is used but as eSata seems to be the future we'll develop software to work over that one too.
randy allen
"The NY FAB will be the most advanced manufacturing facility for micro-processors in the world thanks to AMD's revolutionary APM technologies."
Any FAB that will be ready in 4+ years is lightyears ahead of whatever anyone has today.
"Intel has no quad core, they have dual dual-core CPUs."
I don't care if all they have are tiny midgets in a box, if they count numbers faster than competitors CPUs can then I'm happy.
penix
"What I was trying to convey is that HyperTransport is a readily available and proven option, but Intel is developing QuickPath anyway"
Wouldn't you think that having a competitor would make HT people work harder? Competition is good, after all.
"The original FireWire is still a superior performing technology, which continues to best USB after more than a decade"
Performance, yes. Price, hell no. Also has anyone got any ideas how much power does one FW connector take compared to USB? It is kind of important when one connects battery powered stuff together.
viivmPenix said ...
This is reminiscent of the early FireWire vs USB battle. The original FireWire is still a superior performing technology, which continues to best USB after more than a decade. Intel successfully pushed USB adoption and has displaced FireWire as a dominant technology. This is a success for Intel, but unfortunate because the industry suffers.
I guess you are pretty clueless on interface choice and the economical impact on becoming interface of choice.
The original USB was meant to be user friend (plug and play), replacing most of the legacy interafces such as COM port, parrallel port, PS2, etc. You do not need a high bandwidth bus to handle this, and you need to make sure it is cheap for its wide spread adoption (you need overcome the chicken and egg problem).
Then due to its success (which you claimed failure), its usage has expandd to higher bandwidth task, and thus you see the USB2.
as most of the people here has stressed (when their favorite CPU company losed out to the competitor), you do not need a ultrafast computer to do a normal task.
The USB bandwidth is generally more than enough for all its targeted tasks. yoor mouse will not move faster just because use connect it thru the firewire instead of USB. Nor do you priter task has any significant difference.
For any other tasks that requirement bandwidth more than USB can provide, then other interface (such as firewire) should be considered.
I wonder how you claim the industry suffer. There might be some initial cost in designing it, but later are just reapping what has been sown. putting the acceptence of firewire and USB to be equal, the firewire part will be always at higher cost due to its designs. Firewire's initially asking for royalties is a way to not become dominant which is a way to bring the overall cost down too.
a seperate note, yeah, i know, whatever intel did, is wrong in the eyes of the quite some fanbois here, even though they benefited from the effort :)
Ho Ho said...
Pretty much every single motherboard has at least two firewire ports.
I disagree about the availability of FireWire ports. Newegg.com lists 570 motherboards with USB, but only 211 have FireWire. If I were to buy a flash drive today, I would buy USB because less than half the computers out there will be able to use a FireWire based option. If the adoption was equal, I would gladly spend the few extra dollars for the FireWire option because it would have far superior performance.
Ho Ho said...
Advances in production lower costs.
And production doesn't advance without adoption. It's a synergistic cycle.
Ho Ho said...
I can see zero reasons to why would keyboard or mouse need more bandwidth than USB 1.1 provides.
A keyboard and mouse do not. But other devices do need more. If there wasn't a need for more bandwidth, USB 2.0 would never have been created.
Ho Ho said...
If it would be good enough to use for simple external devices as mouse, keyboard, printer, flash drive etc there would exist such items. Why there aren't?
FireWire adoption has taken such a hit that there are few devices on the market these days. FireWire based printers, scanners and flash drives can be found. Mice and keyboards are MIA, but I suspect that they were available at some time.
Ho Ho said...
As I said, eSata will likely replace USB and wirewire as the future technology for connecting external hard drives and other devices needing massive bandwidth.
I agree that it will take the lead for high bandwidth devices, such as external hard drives.
Ho Ho said...
"It is not a generic connection like USB or FireWire"
Technically, it is.
I was unaware that it could be used for other devices, but I still doubt it will be adopted to replace USB and FireWire for generic usage unless it evolves significantly. There are 3 problems I foresee. First, eSATA does not provide power like USB or FireWire. Second, I do not see any eSATA hubs or repeaters on the market yet. Third, there are still compatibility problems that exist with signal degradation at higher speeds. For these reasons I doubt it will ever take over the market space occupied by USB and FireWire.
Ho Ho said...
Wouldn't you think that having a competitor would make HT people work harder? Competition is good, after all.
Competition can drive the market, but monopolistic tactics hinder the benefits of competition. Intel has used these tactics several times in the past, and will most likely do it again. I predict that QuickPath will be the next USB. It will be an inferior product pushed out to the market in lieu of the superior HyperTransport.
penix
"I would gladly spend the few extra dollars for the FireWire option because it would have far superior performance."
You do know that no flash drive is faster than USB 2.0, do you?
"And production doesn't advance without adoption. It's a synergistic cycle."
Yes but mostly better thing gets adopted. Better might not always mean faster or technically better. Often simplicity and price are the most important things. Firewire didn't have either.
"If there wasn't a need for more bandwidth, USB 2.0 would never have been created."
There was a need for more bandwidth than USB 1.1 provided. There are only a few things that need more than USB 2.0 provides.
" Mice and keyboards are MIA, but I suspect that they were available at some time."
I'm quite sure there are no such things. For one thing, firewire needs much bigger wires (shielding) and it would be quite cumbersome to attach a mouse to such a wire.
"I still doubt it will be adopted to replace USB and FireWire for generic usage unless it evolves significantly."
It will certainly not "kill" USB but firewire is under serious threat.
"eSATA does not provide power like USB or FireWire"
Yes but boxes that have multiple drives in them will have external power anyway.
"Third, there are still compatibility problems that exist with signal degradation at higher speeds"
Any links?
"Competition can drive the market, but monopolistic tactics hinder the benefits of competition. Intel has used these tactics several times in the past, and will most likely do it again"
Pretty much any companie would do it, only difference is that few of them have enough recources. Intel is one who does have. Though I still don't understand what did Intel monopolistic tactics have to do with USB adoption over firewire specifically.
"I predict that QuickPath will be the next USB. It will be an inferior product pushed out to the market in lieu of the superior HyperTransport."
I wouldn't be counting it out that soon. From the nearly non-existent information I have I wouldn't be surprised to see it having tiny performance lead over 16bit HT3.
pointer said...
For any other tasks that requirement bandwidth more than USB can provide, then other interface (such as firewire) should be considered. I wonder how you claim the industry suffer.
The obvious fit for this scenario is external hard drives. We know that good old FireWire 400 will greatly outperform USB 2.0 for this purpose. Now go to a store and look at what is on the self. According to NewEgg.com, there are 536 drive enclosures with USB and 109 with FireWire. Motherboards listed are 570 with USB, and 211 FireWire. Cases with front ports, 769 USB, 422 FireWire.
In this given situation, by your logic, FireWire is a better option than USB and should be considered. But look at the poor FireWire ratios and you will see that USB will win they sale purely because of adoption compatibility. Those ratios should be 1:1 and USB 2.0 shouldn't exist. This is how the industry has suffered.
Ho Ho said...
You do know that no flash drive is faster than USB 2.0, do you?
This subject is open to debate. I believe that there is flash memory that will exceed the real world performance of USB 2.0. I've seen flash memory rated as high as 300x, which translates to roughly 45MB/s. USB 2.0 has a theoretical max of 480Mb/s or 60MB/s. Due to the inferior architecture of USB, 60MB/s is never attained. As evil_merlin showed, a real world benchmark resulted in under 15MB/s. If we make an assumption that the 300x card gets at least half of it's theoretical performance (which has been exceed in some benchmarks), that would be 22.5MB/s, a 33% gain over USB 2.0.
I realize I'm talking about flash cards, not drives, but there isn't enough product for comparison. Since USB dominates the flash drive market, you will not see one that is faster than USB 2.0, but the potential exists.
Ho Ho said...
Yes but mostly better thing gets adopted.
If that were the case, NetBurst wouldn't have had the volume of sales it did.
Ho Ho said...
"Third, there are still compatibility problems that exist with signal degradation at higher speeds"
Any links?
If you check out the drive enclosures at NewEgg.com and scroll through some of the reviews, there are many people experiencing problems with flaky connections. The solution is to drop the speed from 3Gb/s to 1.5Gb/s. I do not know the cause of this problem, but regardless it is a problem with no clear cause which spans multiple products.
This comment has been removed by the author.
This comment has been removed by the author.
What moron said "I predict this will be another Intel failure of equal or greater magnitude than USB. "
Oh yeah, that would be Penix, the parrot moron with Sharikou's cock in his mouth and fingers up his ass, balls pulled firmly back, because we all know, he doesn't like being a man.
Go ahead and find one thing innaccurate in my postings. You won't be able to. I provided links back to the pertinant information too. It's not my fault you are too much of a fanboi to get it thru the thick skull you were born with.
Oneexpert, is all you can do cut and past from Wiki?
penix said "I predict this will be another Intel failure of equal or greater magnitude than USB. "
Penix you're a moron of the greatest magnitude!
oneexpert said "BUY AMD energy saving high performance cpus, platforms and video solutions."
Oneexpert you're another moron!
More power to INTEL!
Fuck AMD, BK by Q2'08!
More power to INTEL!
Fuck AMD, BK by Q2'08!
Either Sharikou has gone crazy or someone has gotten ahold of his password!
Penix said ..
The obvious fit for this scenario is external hard drives. We know that good old FireWire 400 will greatly outperform USB 2.0 for this purpose. Now go to a store and look at what is on the self. According to NewEgg.com, there are 536 drive enclosures with USB and 109 with FireWire. Motherboards listed are 570 with USB, and 211 FireWire. Cases with front ports, 769 USB, 422 FireWire.
In this given situation, by your logic, FireWire is a better option than USB and should be considered. But look at the poor FireWire ratios and you will see that USB will win they sale purely because of adoption compatibility. Those ratios should be 1:1 and USB 2.0 shouldn't exist. This is how the industry has suffered.
what kind of logic is that? the wide spread adoption has a reason, beucase the USB2 bandwidth fit the needs. While firewire is possibly faster (i'm not sure), but there won't be much different in harddisk accesses. What I said was you were given a choice, and you can go ahead to buy the necessary part that fit your need.
What kind of joke to blame an success of the a technology for a lack of adoption in the other technology. Silk is definitely better than cotton too. Should everyone wear silk instead of cotton?
Intel already co-developed MII (ethernet media independent interface), why AMD went and codevelop a variant SMII which according to your theory is going to add cost (the initial design phase). I stil think there is place for SMII such as multi-PHY.
Why there is so many busses out there such as SMBUS which has low bandwidth. All of this has a tecnology needed as well as an economical one.
On firewire vs usb end, you need 4 pins vs 2 pins in a chipset to support a port. now multiply with 10 or 12 ports that you are going to see in the current chipsets offerring.
Stop whining when your favorite technology is not being widely adopted, there must be a reason, either a technological one, economical one or both. In the firewire vs USB case, the reason are both. The fittest will survive most of the time. (sometime there is luck element involve, thus i use most of the time)
Just a side note, what the original USB developer never think of, is how the power pins and the interface are going to be 'missuse' as charger/power supply. :) It is really cool to have one interface for charge your handphone, at the same time manage your handphone. And to the greatest extend of missuse, you can see mini fridge ... :)
This comment has been removed by the author.
Just a side note, what the original USB developer never think of, is how the power pins and the interface are going to be 'missuse' as charger/power supply. :) It is really cool to have one interface for charge your handphone, at the same time manage your handphone.
I have no problem to charge my HTC pda phone via my PC USB port, in fact at the same time I also still can surfing internet (thru my phone), make a call to my friends, sending sms, transfer file from my phone to pc or pc to phone, etc.. without any problems.
oneexpert,
I need you to analyse this power results:
Diamond Viper Radeon HD 2900XT 1GB Videocard Review
AMD 6400+ BLACK EDITION REVIEW: http://www.amdzone.com/modules.php?op=modload&name=Sections&file=index&req=viewarticle&artid=316&page=1
Wow, it looks pretty good compared with AMD's other CPUs! Too bad there were no Intel CPUs for comparison. I suppose that's why it looks decent!
I'm with you Sharikou, AMD BK by Q2'08!
This comment has been removed by the author.
AMDZONE reviews are a joke. Only douchebags like Chris Tom can come up with a review like that and only Douchebags like Sharikou, The_Ghost will read and applaud them.
haha
randy allen
"Sharikou has gone crazy"
Wasn't he before? I mean no sane people can make the claims he does.
giant
"AMD 6400+ BLACK EDITION REVIEW"
That they call a benchmark? To what are they comparing the CPUs? OC'd low-end models at relatively low clock speed vs highest speed CPU you can't run with OC in half the benchmarks. Well, AMDzone can't do any half-decent bencmarking, that's for sure. Also who would get that 6400+ when Q6600 is nearly at same price?
evil_merlin said...
Go ahead and find one thing inaccurate in my postings. You won't be able to.
Fine, we'll do it this way. I gave you your shot.
evil_merlin's claim #1:
FireWire performance decreases under load.
Using a benchmark from digit-life.com you posted the following two win2k benchmarks:
USB 2.0 FireWire
We see that USB 2.0 performance remains at a slow ~14MB/s. FireWire's performance peaks at around 40MB/s, then slowly tapers off to under 25MB/s as the transfer progresses. This is where evil_merlin reaches comes to his first conclusion. But wait... he didn't include the 3rd graph:
Direct IDE
Now we see that the direct ide connection follows a very similar pattern to FireWire and performance decreases as the transfer progresses. They are using 40GB drives for the test and transferring a full 40GB. The performance bottleneck is the hard drives as they reach full capacity. It has nothing to with FireWire.
Furthermore, the graph isn't showing load, it's showing progress. Nor would load matter on FireWire because it uses an independent controller from the CPU and load doesn't affect performance.
Claim #1: FireWire performance decreases under load.
Verdict: Completely false.
evil_merlin's claim #2:
FireWire uses more CPU for file transfers at equivalent speeds.
Using the same benchmark we see the following CPU utilization results for the Win2K transfer:
USB 2.0: 5.93%
FireWire: 9.25%
Direct IDE: 11.2%
Clearly USB is the lowest, and FireWire significantly higher. That doesn't make sense. FireWire doesn't even use the host CPU for transfers because it has a controller which runs completely independently. The problem thus far is this is average CPU usage, but doesn't factor in transfer speed. We need to calculate the average transfer rate. The USB rate is consistent, but FireWire drops at a close to linear rate. By averaging the peak and low rates, we can gather an approximate.
USB 2.0: 14MB/s
FireWire: 30.25MB/s [calculated using ((35 + 40) / 2) + 23 ) / 2]
Direct IDE: 32.5MB/s [calculated using (42 + 23) / 2]
Now we can normalize the CPU utilization using USB as the base.
USB 2.0: 5.93%
FireWire: 4.28% [calculated using 14 / 30.25 * 9.25]
Direct IDE: 4.82% [calculated using 14 / 32.5 * 11.2]
Just for kicks, let's do the same for the Win98SE benchmark results they provided.
USB 2.0: 15.7%
FireWire: 5.33%
Direct IDE: 6.28%
Claim #2: FireWire uses more CPU for file transfers at equivalent speeds.
Verdict: Completely false.
evil_merlin, I gave you more than adequate time to revise your claims. Rather than take the opportunity, you have proven to everyone that you are too incompetent to make even a basic analysis. This event can be regarded as no less than a monumental failure. You are an embarrassment to yourself, anyone who was foolish enough to put their faith in you. On top of all this, you are a pathetic coward. I suspect that even now, rather than admit your failure like a man, you will respond as you always do, with a truly base and hollow insult.
penix,
you are wasting your time with that one. He doesn’t even know what he knows. In fact he is just trolling with you.
He is one of those that post here to make sure this blog looks ridiculous.
Sharikou it’s so easy to click on the delete icon...
Some of the top 10 google search result hit on sharikou or scientia blog which is amazing. I’m pretty sure some of them are paid to that because of this.
Looks like poor little Penis is having issues with facts.
1.) Did firewire's performance decrease over time? Yes. Did USB? Nope.
2.)Did firewire use more CPU than USB during file transfer? Yes. Lets go back and use an OS thats 10 years old to prove a point you fucking moron. Next you'll be wanting to test with a 16mb video card or a sub 1.0 GHz processor too. Fuckwit.
So in otherwords, my post was right on the money.
You lose again ass hat.
Aguia get your dick outta Penis's mouth.
One thing i do like is that AMD fanbois only choose to use benchmarks and give links when comparing anything other than processors..........but when it comes to the performance of C2D vs K8 all benchmarks and links are paid pumpers.......pathetic..........cant wait till barcelona turd brown edition releases so we can hear new excuses about why C2D whoops it.......
Evil_merlin is one of the good example to show how bad attitude are those Intel fanbois that coming to this site.. I totally agree with aguia, sharikou you need to do something. You should to follow what scientia and abinstein has done. You must filter those shameful and rude words that comes out from the intelers bad mouths.
You must filter those shameful and rude words that comes out from the intelers bad mouths.
Filter, that takes to much time. Just delete the all post and that’s it, even if there’s 1% of good words/analysis, but I really doubt about that.
Just read last evil merlin post he understood 0% of the penix post. I'm absolutely sure he can read but that’s about it, understand things take a lot of neurons and he doest have any of those in his brain.
Once again, AMD fanbois, and Sharikou lovers show they don't understand simple english and refuse to admit the facts.
In response to evil_merlin:
The claims you posted on August 25 are not the same as your previous. You're not going to be able to pull off any stupid tricks here. You are clearly the most ignorant visitor to this forum and you will be unable to outsmart anyone.
IBM’s upcoming X4 chipset replaces the current X3 chipset and is scalable up to 32 processors across eight system boards for 128 cores in a single server system. Memory support is plentiful with up to 16 memory slots per system board. Memory modules in 1GB, 2GB and 4GB sizes are supported with the X4 chipset, providing a maximum of 512GB of system memory.
32 processor, 128 core. That's, uh, four times more than AMD. Remind me, which is the more scalable one here?
Post a Comment
<< Home