Some give credit to IBM who in 1969 unbundled their software from their hardware. Prior to that, software was provided free with source code. This move was, in part, a response to anti-trust accusations.
If you use a smaller representation, then more numbers (cells) fit into your cache. This will give speed improvements especially if some algorithm is forcing re-calcs regularly.
Intel's "Enhanced Dynamic Acceleration Technology" is a triumph of marketing. Notice how the focus is on the transition where one core becomes inactive and the other one speeds up. This is the good transition. The other transition, where the chip workload increases & voltage/frequency are limited to keep within a power envelope, is called "throttling" and is much disliked in the user community.
Don't get me wrong, this is valuable technology. It is important that microprocessors efficiently use the power available to them. Having a choice on a single chip between a high-performance, high-power single-thread engine & a set of lower-performance, lower-power engines has great promise. But, the way this is presented is a big victory for marketing.
Microsoft has claimed publicly that about half of XBOX360 users have joined XBOX live. No word on how many are regular users vs. tried it once. Also, I believe that Xbox live access is required to run old Xbox 1 games on XBOX 360, so some folks may have joined just for that.
Operating income (profit) information taken from Sony's financial disclosures (on its web site under investor relations):
FY 2004 (ended 31 March 2005)
(millions of dollars, converted from Yen)
Electronics -321
Games 404
Music 82
Pictures 597
Financial Services 519
Other -38
Total 1243
So, games was about 1/3 of Sony's profits for the most recent year. Certainly much more than music over that period.
And, to gauge against XBOX 360 sales, they sold 16.17 million PS2 units between 1 April 2004 and 31 March 2005.
Disclaimer: I worked on the microprocessor in the XBOX 360 and also on the Cell processor expected in PS3. I game on both XBOX 360 and PS2.
Remember that there are two types of testing involved...bringup testing to verify that the design is right and manufacturing testing to verify that a particular chip was built correctly.
Most modern microprocessors contain build-in self test for the manufacturing test of logic, arrays and IO circuits. These are basically test cases (really test case generators) built into the part.
Also there are significant features for tracing & watching what happens on the chip to assist in bringup.
Since I am not involved with IBM's workstation rollout, I cannot offer details. But, I can say this: in my dealings with IBMers the only folks I have seen with desktop machines are administrative assistants. Everyone else (mostly technical folks in CPU design & throughout IBM Research, but also HR people, web-site maintainers, global services project leaders, managers, executives, etc) have laptops as their personal workstations. That includes folks in the US, Europe & Israel. There are many fixed "deskside" machines which are used to control equipment, act as small servers, test software, etc.
I would expect that a large percentage of the 40k "desktops" quoted are really laptop computers.
Now, I have seen some folks with dual-boot laptops who attempt to run Linux most of the time, so ther e is some progress being made there.
"You don't know what you're talking about."
Yes, i do.
I apologize. My remark was offensive and inappropriate.
Having seen more than one medium-sized company deploy desktops, I'm positive that all computers are already configured to run anything they need to run and print anywhere they need to print *before* they are presented to the user.
You don't know what you're talking about. I did a quick search and found about 2500 public printers in IBM's US offices. And that doesn't include printers in private locations, on isolated sub-networks, etc. I would not want my laptop to have come pre-configured for all of these...and have to scroll through them whenever I want to print.
But, when I travel next week with my new laptop, I will be able to look at a printer, read its ID and configure my laptop to print there by downloading & installing the appropriate software from a web form.
This is a base requirement in large companies.
Similarly, the job families within IBM have tremendous variety. We can't provision new laptops with all the software someone might need. In my job I absolutely need software to communicate with my pSeries box using X11. I wouldn't want IBM to have to purchase 320,000 licenses so that everyone else could have yet another program to clutter up their box. Likewise, I don't want to have to deal with all of the configuration/pricing/etc support software that the sales force has or the troubleshooting/etc software that our repair folks need.
In the training sessions I've taken part in (at the IBM TJ Watson Research lab) they give us scripts to read. Sessions take place with various amounts of background noise (sound-proof room, over a telephone, in the cafeteria, in a car).
And they give us a free lunch coupon afterwards. Will read script for food.:-)
ViaVoice is a wide-vocabulary speech recognition. The article hints at more focused set of target words (times, dates, locations) for the donated package.
Sounds much more like the software supporting airlines which use voice recognition systems to help you request flight information.
The strategies are quite different.
ViaVoice encourages you invest some of your time reading training scripts so it can learn your voice and thus recognize a wide variety of words from your specific voice.
The time/date/city system is likely to be speaker independent (no training scripts to read) but much smaller vocabulary.
All 64-bit PowerPCs are 32/64 bit processors running both with no performance penalty.
That's because the 32-bit PowerPC used for many years was a sub-set of the initial 64-bit PowerPC design. This is quite different from the x86 world where an initial 8/16-bit processor has had 32-bit and 64-bit extensions grafted onto it.
Not that there isn't a penalty of some sort for the 32-bit support...lots of "blank upper word if MSR_sf is off" hardware.
But, it's worth it because applications which don't require 64-bit pointers often run faster when compiled in 32-bit mode (less paging, smaller stack, etc). But, we in chip design use plenty of applications that require 64-bit mode.
Viruses work differently...they keep on spreading...or attempting to spread. Law enforcement needs to remind itself of this difference from time to time because "cyber-crime" is a very small percentage of overall crime.
I love this....especially because of the mention of Joshua.
Why? Joshua mentions the moon in his prayer asking for the sun to be stopped.
Joshua 10:12 The day the Lord delivered the Amorites over to the Israelites, Joshua prayed to the Lord before Israel: "O sun, stand still over Gibeon! O moon, over the Valley of Aijalon!"
10:13 The sun stood still and the moon stood motionless while the nation took vengeance on its enemies. The event is recorded in the Scroll of the Upright One. The sun stood motionless in the middle of the sky and did not set for about a full day.
10:14 There has not been a day like it before or since. The Lord obeyed a man, for the Lord fought for Israel!
10:15 Then Joshua and all Israel returned to the camp at Gilgal.
This works for mainframes (and for iSeries boxes) because IBM publishes relative performance numbers for these processors. These can be used as "official" numbers for licensing agreements.
If you tried to use that sort of thing between Power4 boxes & Sparc boxes, then you'd have all sorts of benchmark wars.
This depends on the vendor. Of course, now Power5 is out...two cores with two threads each. Things may change for some vendors.
For example, on a 4-way Power 5 (16 threads) system, IBM charges for 4 copies of AIX.
There are other competing goals besides max performance.
Cost & time-to-market also favor devoting much of the area to cache. Doubling the size of the cache increases the design effort only incrementally, while doubling the number of functional units is much more significant. (think front-end & completion & bypass)
I also was putting "todo" list comments into code back in the early 80's. I had different types of tags to indicate different sorts of concerns.
But, I don't see how this is prior art. How does your grep output automatically back-annotate into the source-code? Claim #1 of the patent includes the language:
in response to completion of a task, modifying the task list during the interactive code development session to indicate that the task has been completed.
And claim #6:
The method of claim 1, further comprising:
associating a location in the source code with the inserted task; and
in response to navigation to the inserted task, displaying a portion of the source code corresponding to the associated location.
Yes, they've also claimed what you do, but that's just SOP when filing for patents.
I disagree. Claim #1 seems to specifically cover only fully computer-implemented methodology covering all steps including the back-annotation into the code. I'll bet that claim was narrowed by the patent office.
The numbers given are most likely micro-ops, but I'm not completely sure.
Most PowerPC instructions map onto one micro-op, so there isn't really much difference if you're counting peak values.
Slashdot Mirror
Some give credit to IBM who in 1969 unbundled their software from their hardware. Prior to that, software was provided free with source code. This move was, in part, a response to anti-trust accusations.
That. You've certainly heard of open-source. Software wants to be free. This means that programmer salaries want to be $0.
Yes.
If you use a smaller representation, then more numbers (cells) fit into your cache. This will give speed improvements especially if some algorithm is forcing re-calcs regularly.
Intel's "Enhanced Dynamic Acceleration Technology" is a triumph of marketing. Notice how the focus is on the transition where one core becomes inactive and the other one speeds up. This is the good transition. The other transition, where the chip workload increases & voltage/frequency are limited to keep within a power envelope, is called "throttling" and is much disliked in the user community.
Don't get me wrong, this is valuable technology. It is important that microprocessors efficiently use the power available to them. Having a choice on a single chip between a high-performance, high-power single-thread engine & a set of lower-performance, lower-power engines has great promise. But, the way this is presented is a big victory for marketing.
Microsoft has claimed publicly that about half of XBOX360 users have joined XBOX live. No word on how many are regular users vs. tried it once. Also, I believe that Xbox live access is required to run old Xbox 1 games on XBOX 360, so some folks may have joined just for that.
FY 2004 (ended 31 March 2005)
(millions of dollars, converted from Yen)
Electronics -321
Games 404
Music 82
Pictures 597
Financial Services 519
Other -38
Total 1243
So, games was about 1/3 of Sony's profits for the most recent year. Certainly much more than music over that period.
And, to gauge against XBOX 360 sales, they sold 16.17 million PS2 units between 1 April 2004 and 31 March 2005.
Disclaimer: I worked on the microprocessor in the XBOX 360 and also on the Cell processor expected in PS3. I game on both XBOX 360 and PS2.
Remember that there are two types of testing involved...bringup testing to verify that the design is right and manufacturing testing to verify that a particular chip was built correctly. Most modern microprocessors contain build-in self test for the manufacturing test of logic, arrays and IO circuits. These are basically test cases (really test case generators) built into the part. Also there are significant features for tracing & watching what happens on the chip to assist in bringup.
Did you notice the license plate at about 3:49 in the Our Colony video? Looks like PS3KO to me. Just a subtle dig at the expected competition.
I would expect that a large percentage of the 40k "desktops" quoted are really laptop computers.
Now, I have seen some folks with dual-boot laptops who attempt to run Linux most of the time, so ther e is some progress being made there.
I apologize. My remark was offensive and inappropriate.
I'm positive that all computers are already configured to run
anything they need to run and print anywhere they need to print
*before* they are presented to the user.
You don't know what you're talking about. I did a quick search and found about 2500 public printers in IBM's US offices. And that doesn't include printers in private locations, on isolated sub-networks, etc. I would not want my laptop to have come pre-configured for all of these...and have to scroll through them whenever I want to print.
But, when I travel next week with my new laptop, I will be able to look at a printer, read its ID and configure my laptop to print there by downloading & installing the appropriate software from a web form.
This is a base requirement in large companies.
Similarly, the job families within IBM have tremendous variety. We can't provision new laptops with all the software someone might need. In my job I absolutely need software to communicate with my pSeries box using X11. I wouldn't want IBM to have to purchase 320,000 licenses so that everyone else could have yet another program to clutter up their box. Likewise, I don't want to have to deal with all of the configuration/pricing/etc support software that the sales force has or the troubleshooting/etc software that our repair folks need.
And they give us a free lunch coupon afterwards. Will read script for food. :-)
ViaVoice is a wide-vocabulary speech recognition. The article hints at more focused set of target words (times, dates, locations) for the donated package. Sounds much more like the software supporting airlines which use voice recognition systems to help you request flight information.
The strategies are quite different.
ViaVoice encourages you invest some of your time reading training scripts so it can learn your voice and thus recognize a wide variety of words from your specific voice.
The time/date/city system is likely to be speaker independent (no training scripts to read) but much smaller vocabulary.
That's because the 32-bit PowerPC used for many years was a sub-set of the initial 64-bit PowerPC design. This is quite different from the x86 world where an initial 8/16-bit processor has had 32-bit and 64-bit extensions grafted onto it.
Not that there isn't a penalty of some sort for the 32-bit support...lots of "blank upper word if MSR_sf is off" hardware.
But, it's worth it because applications which don't require 64-bit pointers often run faster when compiled in 32-bit mode (less paging, smaller stack, etc). But, we in chip design use plenty of applications that require 64-bit mode.
Combine that with the liners that are necessary for Cu and you see that the resistance increases more than proportionally.
Viruses work differently...they keep on spreading...or attempting to spread. Law enforcement needs to remind itself of this difference from time to time because "cyber-crime" is a very small percentage of overall crime.
Why? Joshua mentions the moon in his prayer asking for the sun to be stopped.
Taken from the NET bible: (New English Translation) http://bible.org/cgi-bin/netbible.pl?header=on&boo k=jos&chapter=10
If you tried to use that sort of thing between Power4 boxes & Sparc boxes, then you'd have all sorts of benchmark wars.
This depends on the vendor. Of course, now Power5 is out...two cores with two threads each. Things may change for some vendors. For example, on a 4-way Power 5 (16 threads) system, IBM charges for 4 copies of AIX.
There are other competing goals besides max performance.
Cost & time-to-market also favor devoting much of the area to cache. Doubling the size of the cache increases the design effort only incrementally, while doubling the number of functional units is much more significant. (think front-end & completion & bypass)
But, I don't see how this is prior art. How does your grep output automatically back-annotate into the source-code? Claim #1 of the patent includes the language:
And claim #6:
I disagree. Claim #1 seems to specifically cover only fully computer-implemented methodology covering all steps including the back-annotation into the code. I'll bet that claim was narrowed by the patent office.
Modern zSeries boxes suppport IEEE float.
I'm surprised they left out the Sony/Toshiba/IBM partnership developing "Cell"....they mentioned it earlier.
The numbers given are most likely micro-ops, but I'm not completely sure. Most PowerPC instructions map onto one micro-op, so there isn't really much difference if you're counting peak values.