Additionally, the speed improvement of ATA133 vs. ATA100 is _only_ on the read datapath. In other words, the reads occur at 133 vs. 100. The driver still has to do the same number of Programmed I/O ("PIO") cycles to set up a data transfer; the writes still occur at roughly ATA88 speed. So, as the read speed increases, you're improving an ever-increasingly insignificant portion of the pie. At that point this is merely marketing bullet-worthy.:)
I've run ELK's Linux on my 8088-based IBM Portable PC machine. (0.3 Bogomips, WOW!):) That's with 512k of memory. I can run 'vi' on 4 different virtual terminals, that's about it.:)
Additionally, there are patents on performing certain algorithms, etc., which preclude GCC from optimizing in certain manners. Search Slashdot for similar GCC topics, this patent aspect is discussed many times.
The ICH doesn't put ATA transactions on the PCI bus. There's a separate bus between the MCH and ICH, over which the USB, AC'97, IDE, etc. transactions go. So the PCI throughput is independent of the IDE throughput (to a certain extent--the PCI traffic headed to memory/CPU does go over this MCH-ICH link though.) But the maximum throughput over this link is greater than the maximum throughput over PCI alone.
For ATA133, writes can stream off of one drive at 133MB/s (if the chipset can support that...), with two drives you must time-division multiplex that (with significant CPU overhead, well, significant in relationship to the data transferred), but writes max out at around 88MB/s. (ATA100 and beyond were marketing ploys, as they're talking about reads only at that point.) The real issue is this: SCSI hosts and drives support tagged-command queuing, which ATA controllers/drives do not (generally.) That, coupled with CPU controlling overhead, and the low number of drives you can connect per controller, and the few types of devices supported, is what tends to sink ATA when compared to SCSI...:)
Most BIOSes (excluding recent AMD motherboards using the 760MP chipset, as their southbridge's USB 1.1 is broken) implement a USB software stack providing 'legacy' support (read: makes software think it's the old keyboard controller at the usual IO addresses/interrupt) until the OS boots loads its own USB 1.1/2.0 stack and enumerates the devices. It's one reason why the flash chips housing the "CMOS" programming keeps getting bigger: 4MB, then 8MB,...
ATA won't keep going up, and even if they did play with the electrical signalling to accomplish "ATA600", we're past the point of "diminshing returns": the set up of ATA cycles still required CPU-initiated Programed I/O "PIO" cycles, the write speed is capped at around 88MB/s, and the actual read data transfers are only at the 133MB/s. So, you're speeding up only one part of an overall ATA cycle, and as you speed that part up, it's overall significance drops. Hence the cheap-drive-industry move for Serial ATA (new electricals/drives, but using legacy software IDE drivers.)
You have to be careful when comparing USB2 drives though, to other interfaces. Early USB2 drives use what's called a tailgate, which translates USB2 into ATA commands. This becomes the bottleneck, as there isn't much buffering provided, etc. in performing this translation. Native USB2 drives (those which have a discrete chip speaking USB2) will be faster.
No, that's not even guarranteed, you can snoop the EMF emitted by the monitor etc.:) Maybe if you're in a grounded cage, AND you're not connected, etc. then you could be secure.:)
I'd say forget the separate 'net, and just use crypto. What's the problem with that?
I would agree, but the thing with Mozilla is, you
can contribute to the Effort and add features like
these. You cannot do the same with Opera and IE (that I know of at least.) So while it may lag in features now, you can certainly expect their implementation in the future.
There's something to be said for having access to the source, especially if you're in an IT deptartment and need to customize disk layout and so on.
I downloaded and installed Mozilla 0.9.3, and tried
going to Bank of America's online accounts site, and lo-and-behold, they don't recognize 'Mozilla' as being the proper version of Netscape or IE. So, does anyone know how to make Mozilla masquerade as IE or Netscape 4.x?
DSN: Deep Space Network [of dishes]
Many groups (SETI, etc.) are studying how to pull data out of noisy radio streams, these deep probes provide a good needle in the haystack, if you will.
Slashdot Mirror
Additionally, the speed improvement of ATA133 vs. ATA100 is _only_ on the read datapath. In other words, the reads occur at 133 vs. 100. The driver still has to do the same number of Programmed I/O ("PIO") cycles to set up a data transfer; the writes still occur at roughly ATA88 speed. So, as the read speed increases, you're improving an ever-increasingly insignificant portion of the pie. At that point this is merely marketing bullet-worthy. :)
I've run ELK's Linux on my 8088-based IBM Portable PC machine. (0.3 Bogomips, WOW!) :) That's with 512k of memory. I can run 'vi' on 4 different virtual terminals, that's about it. :)
Additionally, there are patents on performing certain algorithms, etc., which preclude GCC from optimizing in certain manners. Search Slashdot for similar GCC topics, this patent aspect is discussed many times.
Thanks again! I'm at my DSL max (~180kiB/s.) :)
The ICH doesn't put ATA transactions on the PCI bus. There's a separate bus between the MCH and ICH, over which the USB, AC'97, IDE, etc. transactions go. So the PCI throughput is independent of the IDE throughput (to a certain extent--the PCI traffic headed to memory/CPU does go over this MCH-ICH link though.) But the maximum throughput over this link is greater than the maximum throughput over PCI alone.
The ICH4 southbridge will be out over the summer or the fall this year. It'll have USB2.0 (_much_ faster than NEC's offering), better AC'97, etc.
For ATA133, writes can stream off of one drive at 133MB/s (if the chipset can support that...), with two drives you must time-division multiplex that (with significant CPU overhead, well, significant in relationship to the data transferred), but writes max out at around 88MB/s. (ATA100 and beyond were marketing ploys, as they're talking about reads only at that point.) The real issue is this: SCSI hosts and drives support tagged-command queuing, which ATA controllers/drives do not (generally.) That, coupled with CPU controlling overhead, and the low number of drives you can connect per controller, and the few types of devices supported, is what tends to sink ATA when compared to SCSI... :)
Most BIOSes (excluding recent AMD motherboards using the 760MP chipset, as their southbridge's USB 1.1 is broken) implement a USB software stack providing 'legacy' support (read: makes software think it's the old keyboard controller at the usual IO addresses/interrupt) until the OS boots loads its own USB 1.1/2.0 stack and enumerates the devices. It's one reason why the flash chips housing the "CMOS" programming keeps getting bigger: 4MB, then 8MB, ...
ATA won't keep going up, and even if they did play with the electrical signalling to accomplish "ATA600", we're past the point of "diminshing returns": the set up of ATA cycles still required CPU-initiated Programed I/O "PIO" cycles, the write speed is capped at around 88MB/s, and the actual read data transfers are only at the 133MB/s. So, you're speeding up only one part of an overall ATA cycle, and as you speed that part up, it's overall significance drops. Hence the cheap-drive-industry move for Serial ATA (new electricals/drives, but using legacy software IDE drivers.)
You have to be careful when comparing USB2 drives though, to other interfaces. Early USB2 drives use what's called a tailgate, which translates USB2 into ATA commands. This becomes the bottleneck, as there isn't much buffering provided, etc. in performing this translation. Native USB2 drives (those which have a discrete chip speaking USB2) will be faster.
No, that's not even guarranteed, you can snoop the EMF emitted by the monitor etc. :) Maybe if you're in a grounded cage, AND you're not connected, etc. then you could be secure. :)
I'd say forget the separate 'net, and just use crypto. What's the problem with that?
I would agree, but the thing with Mozilla is, you can contribute to the Effort and add features like these. You cannot do the same with Opera and IE (that I know of at least.) So while it may lag in features now, you can certainly expect their implementation in the future. There's something to be said for having access to the source, especially if you're in an IT deptartment and need to customize disk layout and so on.
I downloaded and installed Mozilla 0.9.3, and tried going to Bank of America's online accounts site, and lo-and-behold, they don't recognize 'Mozilla' as being the proper version of Netscape or IE. So, does anyone know how to make Mozilla masquerade as IE or Netscape 4.x?
set prompt="%B%T %m %l %h%b %~%B %#%b "
It looks like this:
11:29 fhi2009 ttyr7 209 ~ >
It shows 24hr time, machine, terminal, command number, path.
DSN: Deep Space Network [of dishes] Many groups (SETI, etc.) are studying how to pull data out of noisy radio streams, these deep probes provide a good needle in the haystack, if you will.