The sad thing is, in Microsoft's eye, I'd be willing to bet that YOU are the customer, not your users. Therefore, your "few thousand NT users" count as a single customer in their eyes..
Congressman Darrell Issa (R.-Calif.) Wednesday introduced a bill based on a letter to the Pentagon, the U.S. Agency for International Development (USAID) and other lawmakers demanding that the Department of Defense and USAID show favor to CDMA technology made by San Diego-based QUALCOMM.
His constituency (aka Qualcomm) should be bidding on and receiving their reconstruction contracts based on their merit, cost, and feasibility, not because they paid for some yahoo to get elected and that yahoo promptly legislated use of their products. Such a bill would create a false monopoly for Qualcomm without regard to neighboring mobile phone standards and the merits of their competitors' products.
This is the epitome of pork-barrel politics. Senators should not be for sale. This Issa guy obviously is. Shame on him!
It's considerably faster than 56k/v90 whatever; the modem tech is really just 33.6 line speed with enhanced compression, AFAIK.
The V.90 (and V.92) connection speed you see is actually the uncompressed downstream data rate. Separately, the protocol is called V.PCM, as in Pulse Code Modulation, the method the telcos use to carry the raw circuit-switched speech path. When you get a connection speed of 52kbps, that's true 52kbps without compression. Even though it's called a 56k modem, 56k is the theoretical maximum and is never reached due to FCC regulations.
Upstream is a different matter. With V.90, upstream tops out at 33.6k (also uncompressed). I've heard that the V.92 upstream speed is higher, though I don't know how.
FYI, I'm one of the dying breed of telco equipment manufacturer engineers, and tested the V.90 modems with our equipment when they first came out.
Remember that the early TiVo's used a 33MHz processor.
The early TiVo's had an external MPEG decoder chip as well. If the 600MHz Via processor can handle MPEG decoding, as you indicate it can, that's fantastic. However, I believe a generic 33MHz processor would have a really tough time decoding an MPEG stream with any quality. The hardware MPEG decoder boards didn't go away until processors exceeded 400MHz or so..
The power consumption of AMD's mobile processor is still much much higher than Intel's. Tom's Hardware says here that the power-saving features of the Pentium M are supposed to ensure that Pentium-M has an "average power consumption" of less than 1 W, while still delivering satisfying performance. PCWorld corroborates that here stating that the 1.3-GHz, 1.4GHz, 1.5-GHz, and 1.6-GHz Pentium M chips draw an average of less than 1 watt of power.
Compare that to the advertized draw of AMD's low-voltage chips including the 1800+, 1700+, 1600+, 1500+, and 1400+ models which dissipate 25 watts when operating at maximum power. If that's the maximum draw, the average is not likely to be less than 10..
The caveat is that the other laptop conponents, most notably the backlit display, consume the lion's share of the battery life anyway. Lord knows I support the underdog (I even bought a Cyrix instead of an original Pentium), but this Centrino chip is good.. damn good.
So what's to stop someone from scratching off the ticket and just bringing it straight to the local gas station? The most they can do is tell you it isn't a winner. That way you wouldn't have to install their bloated adware on your already unstable copy of Windows 98 just to recover $3 of the $33 you already spent on the stupid things...
On the second point, even if you were to put in serial numbers until you found the $25,000 winner, then what? You still don't have the ticket you'd need to turn into the local gas station. And you can bet they're going to examine the "big winnner" very carefully to make sure it's authentic!
I replaced my floppy drives with LS-120 Superdrives several years ago. The latest PC I built doesn't even have that; just a CD-RW and a NIC. I'd rather the MB makers eliminate the floppy interface than the PCI.. Does anything still come with a floppy? I haven't seen a new one in ages.
Not to mention the security aspect of it.. With Deep Blue, you didn't have to worry about something like this:
Imagine you're playing chess with this "grid player." Only one of the nodes on this grid is a hacked client that is sending bad (but still authentic) results.. Because of this, you advance a pawn instead of moving your knight, leading to your defeat.
Now, imagine this was more life-threatening than a simple game of chess. . . *shudder*
Or...
What if you have a VM that just pretends its TCPA capable, but the TCPA hardware is just emulated by the host O/S's VM software. Now, boot up Palladium within the VM and run a "protected" program. You should have a way of accessing protected memory in the guest O/S simply by reading it from within the host...
Or, looking at it from a different angle, when x86-64 is out, and you can run 32-bit programs within a 64-bit O/S, do you suppose one of those 32-bit "probrams" could be Palladium?
BTW, I realize the whole WVMS within a WVMS thread is just ramblings, but my understanding of how VMs function indicates that that would not be possible. IIRC, a 486-class or better host processor is required to host a VM, and it only virtualizes a 386-class processor.
I'm sorry. Comments of this nature are not appropriate for this forum. This thread is strictly for discussing duplicate stories.
If you wish to discuss the contents of the TomsHardware CPU review, please post your comments in this discussion.
Timothy, you wanna take a crack at it too? ChrisD? Any other/. editors? Bueller? Bueller?
If you have some way of synthesizing tones, you can create your own version of the SIT tone. The one I got from the net was of poor quality and didn't have totally accurate frequencies. This site has a description of the tones and durations and what they indicate to the calling party. IMHO, the best combo is
Vacant Circuit: 985.1Hz for 380ms, 1370.6Hz for 274ms, and 1776.7Hz for 380ms
I heard that the Telezapper just sends the first tone, which is enough for the autodialers to recognize it as some sort of telco announcement. You must answer the phone fairly quickly though, because some dialers assume that the tone will come within a couple seconds after dialing is complete. If you take too long, it may already be connected to the call handler that's going to make the sales pitch.
Oh, yeah. Not only that, but with the memory controller hanging off of the FSB (ala P4 & Athlon XP), all memory access has additional latency as the request traverses the external FSB. By integrating the memory controller, that bus is all on-chip and lightning fast.
Intel will soon be moving to a 800 mhz FSB, while AMD's brand spanking new chip will be stuck at 333.
I'm not buying that argument. I think you're forgetting that the FSB on current PCs needs to be higher than the memory bandwidth specifically because it will be transporting all memory accesses (DMA excluded). While a P4 FSB is effectively 533, it takes dual-channel RDRAM to completely fill that pipe. And the FSB must be shared among ALL things the CPU is accessing, including memory, PCI devices, and AGP.
The Athlon-64 and Opteron use a completely different FSB technology called HyperTransport. The bandwidth is smokin' (I forgot the exact figures atm, but I believe its in the 800-1.2G range), and it doesn't have to carry memory accesses, since that's built in. But, as previously mentioned, since each processor comes with its own crossbar memory controller, you're scaling memory bandwidth each time you scale processing power. On a 2-way system, if processor A needs to access memory B, it sends an HT request directly to B's memory controller and A keeps on chewing on whatever thread its on.
From his article (and FWIW, I am in total agreement): ... it's a total pain in the ass to use due to rampant "themeing." Why do people do this? They map this stupid shaped window with no titlebar (oh, sorry, your choice of a dozen stupidly-shaped windows without titlebars) all of which use fonts that are way too small to read.
One of the reasons M$ Windows has done so well is that it looks the same from one machine to another and from one program to another. If I have Windows on my computer, I know how to use your Windows computer. If you know how to use Word, you know how to use Excel. The menus are in the same order and have largely the same items. The active titlebar is a different color from the inactive ones, and clicking on it raises that window to the top. Standard, default appearances and actions! How WinXP is turning it into a Fischer-Price toy is a rant for a different day..
One of the most well liked "themed" programs for Windows is WinAMP. I submit to you that one of the reasons it was so well accepted is that the default skin looked like a normal window! Only the color and size were different. That meant that my mom (BTW, she still can't spell WWW..) knew how to resize it, move it and close it.. Instead of having a round volume control, like a home audio componenet, it had a slider bar, like a *gasp* PROGRAM! (Clue for those that need one: WinAMP is a program.)
Developers, if you want to give your interface themes and skins and other "fluff", by all means, knock yourself out. However, the default skin should be one that implements the interface as it would appear without a skin. Please! For everyone out there how likes to make their computer look like Fantasia, there are probably more of us who like it to look like a computer.
I'm hoping the FPGA will be a resource that can be allocated just like memory. That would make it a great solution for implementing DSP-based . A media player could reserve the FPGA, load the MP3 codec into the FPGA and start feeding it the compressed bitstream. When a different bitstream comes along (let's say Ogg?), clear out the current codec and pop in the new one. Meanwhile, your processor load stays close to zero because all the number-crunching is going on in programmable hardware!
DRM software could use the FPGA in conjunction with the TCPA chip for access control. However, the TCPA chip has much less nefarious uses as well, such as hardware encryption/decryption and secure key storage to name a couple. If a newer, better encryption algorithm comes along, it could be implemented in the FPGA, making this platform extremely future-proof.
I wish IBM the best in this. It sounds like a truly marvelous platform.
Speaking of orthogonal instruction sets (ok, I'm stretching here), there's a series of x86 instructions to exchange another register with AX.
93 exchanges DX and AX.
92 exchanges CX and AX.
91 exchanges BX and AX.
90 is listed as NOP.
No operation? It actually does give the processor something to do: exchange AX with itself!
A MOV with an immediate doesn't pull anything over the memory bus other than the instruction and operands which any operation would. However, it would need to move the immediate from the CPU's instruction cache into the ALU.
Its still worse than XORing a register with itself, but its not as bad as you make it out to be. Really a moot point, being that XOR AX, AX is a better choice overall and MOV AX, #0 has absolutely no advantages.
Slashdot Mirror
I support a few thousand NT users at a plant...
The sad thing is, in Microsoft's eye, I'd be willing to bet that YOU are the customer, not your users. Therefore, your "few thousand NT users" count as a single customer in their eyes..
Congressman Darrell Issa (R.-Calif.) Wednesday introduced a bill based on a letter to the Pentagon, the U.S. Agency for International Development (USAID) and other lawmakers demanding that the Department of Defense and USAID show favor to CDMA technology made by San Diego-based QUALCOMM.
His constituency (aka Qualcomm) should be bidding on and receiving their reconstruction contracts based on their merit, cost, and feasibility, not because they paid for some yahoo to get elected and that yahoo promptly legislated use of their products. Such a bill would create a false monopoly for Qualcomm without regard to neighboring mobile phone standards and the merits of their competitors' products.
This is the epitome of pork-barrel politics. Senators should not be for sale. This Issa guy obviously is. Shame on him!
It's considerably faster than 56k/v90 whatever; the modem tech is really just 33.6 line speed with enhanced compression, AFAIK.
The V.90 (and V.92) connection speed you see is actually the uncompressed downstream data rate. Separately, the protocol is called V.PCM, as in Pulse Code Modulation, the method the telcos use to carry the raw circuit-switched speech path. When you get a connection speed of 52kbps, that's true 52kbps without compression. Even though it's called a 56k modem, 56k is the theoretical maximum and is never reached due to FCC regulations.
Upstream is a different matter. With V.90, upstream tops out at 33.6k (also uncompressed). I've heard that the V.92 upstream speed is higher, though I don't know how.
FYI, I'm one of the dying breed of telco equipment manufacturer engineers, and tested the V.90 modems with our equipment when they first came out.
Remember that the early TiVo's used a 33MHz processor.
The early TiVo's had an external MPEG decoder chip as well. If the 600MHz Via processor can handle MPEG decoding, as you indicate it can, that's fantastic. However, I believe a generic 33MHz processor would have a really tough time decoding an MPEG stream with any quality. The hardware MPEG decoder boards didn't go away until processors exceeded 400MHz or so..
On the dusty floor is a gold pocket watch. [East] [West] [North] [South]
Good post.
Apparently, I got my specs mixed up. Thanks for setting us straight.
I'm glad to see AMD isn't as far behind as I thought at first glance..
The power consumption of AMD's mobile processor is still much much higher than Intel's. Tom's Hardware says here that the power-saving features of the Pentium M are supposed to ensure that Pentium-M has an "average power consumption" of less than 1 W, while still delivering satisfying performance. PCWorld corroborates that here stating that the 1.3-GHz, 1.4GHz, 1.5-GHz, and 1.6-GHz Pentium M chips draw an average of less than 1 watt of power.
Compare that to the advertized draw of AMD's low-voltage chips including the 1800+, 1700+, 1600+, 1500+, and 1400+ models which dissipate 25 watts when operating at maximum power. If that's the maximum draw, the average is not likely to be less than 10..
The caveat is that the other laptop conponents, most notably the backlit display, consume the lion's share of the battery life anyway. Lord knows I support the underdog (I even bought a Cyrix instead of an original Pentium), but this Centrino chip is good.. damn good.
So what's to stop someone from scratching off the ticket and just bringing it straight to the local gas station? The most they can do is tell you it isn't a winner. That way you wouldn't have to install their bloated adware on your already unstable copy of Windows 98 just to recover $3 of the $33 you already spent on the stupid things...
On the second point, even if you were to put in serial numbers until you found the $25,000 winner, then what? You still don't have the ticket you'd need to turn into the local gas station. And you can bet they're going to examine the "big winnner" very carefully to make sure it's authentic!
I replaced my floppy drives with LS-120 Superdrives several years ago. The latest PC I built doesn't even have that; just a CD-RW and a NIC. I'd rather the MB makers eliminate the floppy interface than the PCI.. Does anything still come with a floppy? I haven't seen a new one in ages.
Vat ees thees "fla-pee" you speek off?
Not to mention the security aspect of it.. With Deep Blue, you didn't have to worry about something like this:
Imagine you're playing chess with this "grid player." Only one of the nodes on this grid is a hacked client that is sending bad (but still authentic) results.. Because of this, you advance a pawn instead of moving your knight, leading to your defeat.
Now, imagine this was more life-threatening than a simple game of chess. . . *shudder*
...Its free-software counterpart is plex86....
That was before Plex86 was "simplified to be a user (application) code only Virtual Machine technology" and became much more UML-like..
Or...
What if you have a VM that just pretends its TCPA capable, but the TCPA hardware is just emulated by the host O/S's VM software. Now, boot up Palladium within the VM and run a "protected" program. You should have a way of accessing protected memory in the guest O/S simply by reading it from within the host...
Or, looking at it from a different angle, when x86-64 is out, and you can run 32-bit programs within a 64-bit O/S, do you suppose one of those 32-bit "probrams" could be Palladium?
BTW, I realize the whole WVMS within a WVMS thread is just ramblings, but my understanding of how VMs function indicates that that would not be possible. IIRC, a 486-class or better host processor is required to host a VM, and it only virtualizes a 386-class processor.
I'm sorry. Comments of this nature are not appropriate for this forum. This thread is strictly for discussing duplicate stories.
/. editors? Bueller? Bueller?
If you wish to discuss the contents of the TomsHardware CPU review, please post your comments in this discussion.
Timothy, you wanna take a crack at it too? ChrisD? Any other
Is "hypercopmuter" a real word with a standardized definition?
Never heard of it. But anything to quiet those pesky, over-zealous, redneck sheriff's deputies sounds good to me!
Sorry, couldn't pass it up...
If you have some way of synthesizing tones, you can create your own version of the SIT tone. The one I got from the net was of poor quality and didn't have totally accurate frequencies. This site has a description of the tones and durations and what they indicate to the calling party. IMHO, the best combo is
Vacant Circuit: 985.1Hz for 380ms, 1370.6Hz for 274ms, and 1776.7Hz for 380ms
I heard that the Telezapper just sends the first tone, which is enough for the autodialers to recognize it as some sort of telco announcement. You must answer the phone fairly quickly though, because some dialers assume that the tone will come within a couple seconds after dialing is complete. If you take too long, it may already be connected to the call handler that's going to make the sales pitch.
To the average VB Developer, the words "Obfuscated C" is redundant.
Ok, since we're trolling...
To the average C software engineer, the acronym BASIC clarifies the VB Developer's skill level.
Your statements are just slightly out of order and incomplete. I'll take the liberty of correcting them for you:
... nice box, nice gradient buttons, stylish consistent GUI for a reasonable price? ... For now, it's Fisher Price.
Oh, yeah. Not only that, but with the memory controller hanging off of the FSB (ala P4 & Athlon XP), all memory access has additional latency as the request traverses the external FSB. By integrating the memory controller, that bus is all on-chip and lightning fast.
Intel will soon be moving to a 800 mhz FSB, while AMD's brand spanking new chip will be stuck at 333.
I'm not buying that argument. I think you're forgetting that the FSB on current PCs needs to be higher than the memory bandwidth specifically because it will be transporting all memory accesses (DMA excluded). While a P4 FSB is effectively 533, it takes dual-channel RDRAM to completely fill that pipe. And the FSB must be shared among ALL things the CPU is accessing, including memory, PCI devices, and AGP.
The Athlon-64 and Opteron use a completely different FSB technology called HyperTransport. The bandwidth is smokin' (I forgot the exact figures atm, but I believe its in the 800-1.2G range), and it doesn't have to carry memory accesses, since that's built in. But, as previously mentioned, since each processor comes with its own crossbar memory controller, you're scaling memory bandwidth each time you scale processing power. On a 2-way system, if processor A needs to access memory B, it sends an HT request directly to B's memory controller and A keeps on chewing on whatever thread its on.
Ok, I know its off topic and all, but mod the parent up funny!
And here's a pullet surprise winning poem too accompany it:
I have a spelling checker.
It came with my PC.
It plainly marks four my revue
Mistakes I cannot sea.
I've run this peom threw it.
I'm sure your pleased two no.
Its letter perfect in it's weigh;
My checker tolled me sew.
Ok, here's some constructive criticism:
... it's a total pain in the ass to use due to rampant "themeing." Why do people do this? They map this stupid shaped window with no titlebar (oh, sorry, your choice of a dozen stupidly-shaped windows without titlebars) all of which use fonts that are way too small to read.
From his article (and FWIW, I am in total agreement):
One of the reasons M$ Windows has done so well is that it looks the same from one machine to another and from one program to another. If I have Windows on my computer, I know how to use your Windows computer. If you know how to use Word, you know how to use Excel. The menus are in the same order and have largely the same items. The active titlebar is a different color from the inactive ones, and clicking on it raises that window to the top. Standard, default appearances and actions! How WinXP is turning it into a Fischer-Price toy is a rant for a different day..
One of the most well liked "themed" programs for Windows is WinAMP. I submit to you that one of the reasons it was so well accepted is that the default skin looked like a normal window! Only the color and size were different. That meant that my mom (BTW, she still can't spell WWW..) knew how to resize it, move it and close it.. Instead of having a round volume control, like a home audio componenet, it had a slider bar, like a *gasp* PROGRAM! (Clue for those that need one: WinAMP is a program.)
Developers, if you want to give your interface themes and skins and other "fluff", by all means, knock yourself out. However, the default skin should be one that implements the interface as it would appear without a skin. Please! For everyone out there how likes to make their computer look like Fantasia, there are probably more of us who like it to look like a computer.
</rant> Call me old fashioned...
I'm hoping the FPGA will be a resource that can be allocated just like memory. That would make it a great solution for implementing DSP-based . A media player could reserve the FPGA, load the MP3 codec into the FPGA and start feeding it the compressed bitstream. When a different bitstream comes along (let's say Ogg?), clear out the current codec and pop in the new one. Meanwhile, your processor load stays close to zero because all the number-crunching is going on in programmable hardware!
DRM software could use the FPGA in conjunction with the TCPA chip for access control. However, the TCPA chip has much less nefarious uses as well, such as hardware encryption/decryption and secure key storage to name a couple. If a newer, better encryption algorithm comes along, it could be implemented in the FPGA, making this platform extremely future-proof.
I wish IBM the best in this. It sounds like a truly marvelous platform.
Speaking of orthogonal instruction sets (ok, I'm stretching here), there's a series of x86 instructions to exchange another register with AX.
93 exchanges DX and AX.
92 exchanges CX and AX.
91 exchanges BX and AX.
90 is listed as NOP.
No operation? It actually does give the processor something to do: exchange AX with itself!
When I hear the phrase "XOR trick," I think of its usefulness for exchanging registers.
If you don't have an XCHG op like the Motorola 68k to exchange two registers, you can use the following trick:
XOR AX, BX
XOR BX, AX
XOR AX, BX
Now, that's a real trick!
A MOV with an immediate doesn't pull anything over the memory bus other than the instruction and operands which any operation would. However, it would need to move the immediate from the CPU's instruction cache into the ALU.
Its still worse than XORing a register with itself, but its not as bad as you make it out to be. Really a moot point, being that XOR AX, AX is a better choice overall and MOV AX, #0 has absolutely no advantages.