Oh, and there are 31 registers - X0 through X30. The 32nd register is special depending on the instruction - for ADD and SUB, "X31" means the stack pointer. For most other instructions, it means the zero register (reads as zero), something borrowed from MIPS, and allowing interesting register-only instruction forms to be used when the immediate value is zero. It does result in oddball uses though, like
SUB SP, 0, X0 ; Set SP.
to play with the stack pointer.
Hi, since you seem to be familiar with AArch64, perhaps you could explain why X31 == 0 is preferrable to X31 == FF's (all 1's). At least with FF's I can do an Increment or Decrement using 2's complement math, as well as bitwise testing for non-zero bits (though CLZ might suffice). With X31 == 0, the use case seems more limited (zero constant can be encoded pretty easily as an IMMEDIATE, and I presume that TEQ/CMP against 0 still exists)?
PS. I'm familiar with ARM32, and somewhat familiar with MIPS-style fixed register constant usage.
However, from the embedded video in this post, it felt like I'm in a 3D VLSI design tool. Maybe in the future we'll be literally designing devices one block at a time.
Either that or the Nazcas have been playing Minecraft all this while....
I ran into the problem of manipulating the 16 bit pointer data and its arithmetic in an 8 bit machine. My solution to this problem of handling 16 bit data, notably pointers, with an 8 bit microprocessor was to implement a nonexistent 16 bit processor in software, interpreter fashion
i.e. Can an IPhone/ITouch app (even a Jailbroken one?) let you communicate with the other 50 IPhone/ITouch users in the train you're on, without paying the cell companies?
Wouldn't bluetooth be the better transport for this?
I don't believe Bluetooth can handle more than 8 devices in a single PAN.
Yes and no.
You can't create an ad hoc network on your iPhone, but you can join one. So if someone has a notebook and creates (seeds) and ad hoc wifi network, then you can connect to that ad hoc network
(This issue has been raised in the iPhone Dev forum, and Apple is aware of it....for what it's worth).
Basically, with a voice channel, the phone is just sending voice frames (we used to call this circuit switching). With VoIP, you'll have to tag each voice frame with a RTP+UDP+IP header, which ends up being larger than the actual voice data itself.
In addition, as you mentioned voice traffic has QoS enabled, while most Internet traffic runs as best effort.
So, in short, having VoIP over 3G is more of a geek's idea of cool rather than being of any real improvement to either voice quality or network efficiency, it's actually consuming more wireless bandwidth in the process (apart from cannibalizing the telco's revenue stream).
Hi, I agree with your sentiments (I've programmed in 8086 and 68HC11), though currently I'm trying to pick up ARM assembly language.
I find that 'gcc -Os' beats my handcrafted assembly due to the fact that the compiler can make a lot of 'short cut' optimizations based on what it know regarding memory address locations, etc. (via PC-relative indirect addressing) that would be difficult to take advantage of in Assembly without making it a non-maintainable hack.
I'm still trying to figure out how to beat gcc, but I don't think it'll be easy, especially not via micro-optimization where I take a C algorithm and reimplement it as given in Assembly.
That is not the only problem nowadays, even processors within a given family may or may not have specific features (VT, for example) disabled. You'd think that there is a conspiracy going on...
It was probaby due to changes to Network Manager and how it detects 3G modem cards. YMMV. I have a Huawei E220 USB modem. It was fixed during Fedora 11, I use it everyday. However, I had to delete the old settings and recreate the modem entry.
Glad to hear you got it working in F12.
My only issue with text files encrypted with gpg is that one of these days, I'd accidentally type the gpg decryption key on the command line instead of when prompted, and the key will show up nicely in the shell history. After the "Arrgh!!" has subsided I'd have to remember to purge the history file, etc.
I've been slowly trying to pick up LyX to replace MS Word for writing papers and technical documents in Comp. Sci.
Currently the biggest impediment in using LyX is the graphics / diagram integration. There is no easy way to edit the diagram from LyX (unlike MS Word where diagrams are embedded); it is a multi-step process where I'd edit the diagram in some other program (Dia is still not polished enough yet), export to EPS, then relink to the EPS in LyX if the filename were different.
The other problem is that EPS output is pretty spotty in most MS programs. Visio's EPS output sucks, to put it mildly, and often cause postscript errors. SVG export is still not well supported in ANY tool I've tried.
I'd be interested to know what kind of apps and formats works best when dealing with diagrams which needs to be cross-platform (Linux & Mac OS X).
I've never managed to get Word's autonumbering to work correctly in a continuously edited technical document despite many attempts to reset / modify the Heading X styles. The auto-formatting feature in MS Word is the worst culprit. This has always been MS Word's problem from the beginning (I'm currently using Office 2003 out of necessity).
I've never tried Office 2007, didn't want to invest any more energy on learning an incompatible interface on an application that does not deal well with technical documents.
LyX (LaTeX) is great for stuff like auto-numbering, but diagrams are a pain since there's no easy way to edit the diagram by double clicking on it, besides the lack of good diagramming tools that export to EPS format. Visio generates lousy EPS (deliberately?)
FrameMaker was the best of the lot. Unfortunately its market share and numerous bugs during the time Adobe bought it over meant that it's no longer a real contender for most people.
Well, unfortunately the ISO booted program segfaults on my PC when I tried to run the flash utility (AMD Phenom X3, 780G chipset). Thankfully it did not affect the drive.
On the website URL, it states that it has been taken offline temporarily as of Jan 19, 2008 8PM CST. Sigh.
I wish the ports system would have a better dependency resolution mechanism. Every time a new port comes out, it'd frequently trigger recompilations of various other dependencies whether it's needed or not when I specify a forced install. Ditto with removing older packages with multiple versions installed, it'd complain about bogus dependencies which do not affect the older version of the package.
Whereas with RHEL (which I'm most familiar with), RPM would trigger a dependency package install only when it is actually needed.
The probably solution? TPM built into your PC. Then you can't complain about not being able to remove the DRM since it's built in. You PC would be just like a console.
Not that it hasn't been tried before..... copy protection was pretty common in the 80's, then it was removed in the 90's due to hue and cry. I see it as a pendulum, we're just swinging the other way again right now, in a new DRM arms race.
We're running native dual-stacked IPv6 in our lab at the university, and my pet peeve is that rouge routers announcing invalid routes will cause the IPv6 enabled systems (mostly Linux, in our case) to be misrouted (no route to host) until the interface is reset (ifdown/ifup). If there's a timeout mechanism associated with router advertisments, then this wouldn't be such a PITA issue.
The other issue has to do with multihoming for BGP routing. AFAIK it's still a 'research problem'.
Prior to XP-SP2, it was possible to run WinXP Pro on my Fujitsu P2120 (256 MB -> 240 MB due to the x86 translator) at a reasonable speed. However, after upgrading to XP-SP2, it now takes almost 5 minutes from power on to the logged-in desktop after all the System Tray icons has stopped loading due to all the disk paging.
Maybe the statement that it worked with 128 MB might have been true prior to XP-SP2, but it's definitely not true now.
I wonder if this meme was planted by spammers to help them bypass Spam filtering (and conditioning readers to intentionally mis-spelt text). I remembered that a lot of strangely spelt spam flooding my Inbox around the same time. Either that or Spammers latched on to this real quick.
If your definition of blackhole routing is correct, then won't XP SP3 make life easier for botnets to launch DDos attacks, since the XP network stack would just drop the blackholed routes? I hope it's not a supposed 'fix' leading to an even greater problem.
> Allow multiple router announcements and behave gracefully. Use the prefix that best matches the destination.
I'm not sure how it could decide on the 'best' prefix to use if it is completely outside of the assigned subnet prefix. That should be the job of the router(s).
Actually what bugs me about IPv6 autoconfigured router prefixes is that it NEVER expires on the client machine (at least, not on Linux).
Someone bringing up a rogue radvd daemon would kill existing routes since it decides rather arbitrarily which router it would use as the default, and it usually ends up being the non-working rogue prefix.
The only workaround is to down the client interface and bring it up again to get a new autoconfigured prefix.
Slashdot Mirror
Oh, and there are 31 registers - X0 through X30. The 32nd register is special depending on the instruction - for ADD and SUB, "X31" means the stack pointer. For most other instructions, it means the zero register (reads as zero), something borrowed from MIPS, and allowing interesting register-only instruction forms to be used when the immediate value is zero. It does result in oddball uses though, like SUB SP, 0, X0 ; Set SP. to play with the stack pointer.
Hi, since you seem to be familiar with AArch64, perhaps you could explain why X31 == 0 is preferrable to X31 == FF's (all 1's). At least with FF's I can do an Increment or Decrement using 2's complement math, as well as bitwise testing for non-zero bits (though CLZ might suffice). With X31 == 0, the use case seems more limited (zero constant can be encoded pretty easily as an IMMEDIATE, and I presume that TEQ/CMP against 0 still exists)? PS. I'm familiar with ARM32, and somewhat familiar with MIPS-style fixed register constant usage.
However, from the embedded video in this post, it felt like I'm in a 3D VLSI design tool. Maybe in the future we'll be literally designing devices one block at a time. Either that or the Nazcas have been playing Minecraft all this while....
I ran into the problem of manipulating the 16 bit pointer data and its arithmetic in an 8 bit machine. My solution to this problem of handling 16 bit data, notably pointers, with an 8 bit microprocessor was to implement a nonexistent 16 bit processor in software, interpreter fashion
Linux on AVR ATMega
i.e. Can an IPhone/ITouch app (even a Jailbroken one?) let you communicate with the other 50 IPhone /ITouch users in the train you're on, without paying the cell companies?
Wouldn't bluetooth be the better transport for this?
I don't believe Bluetooth can handle more than 8 devices in a single PAN.
Yes and no. You can't create an ad hoc network on your iPhone, but you can join one. So if someone has a notebook and creates (seeds) and ad hoc wifi network, then you can connect to that ad hoc network (This issue has been raised in the iPhone Dev forum, and Apple is aware of it....for what it's worth).
Basically, with a voice channel, the phone is just sending voice frames (we used to call this circuit switching). With VoIP, you'll have to tag each voice frame with a RTP+UDP+IP header, which ends up being larger than the actual voice data itself.
In addition, as you mentioned voice traffic has QoS enabled, while most Internet traffic runs as best effort.
So, in short, having VoIP over 3G is more of a geek's idea of cool rather than being of any real improvement to either voice quality or network efficiency, it's actually consuming more wireless bandwidth in the process (apart from cannibalizing the telco's revenue stream).
Hi, I agree with your sentiments (I've programmed in 8086 and 68HC11), though currently I'm trying to pick up ARM assembly language.
I find that 'gcc -Os' beats my handcrafted assembly due to the fact that the compiler can make a lot of 'short cut' optimizations based on what it know regarding memory address locations, etc. (via PC-relative indirect addressing) that would be difficult to take advantage of in Assembly without making it a non-maintainable hack.
I'm still trying to figure out how to beat gcc, but I don't think it'll be easy, especially not via micro-optimization where I take a C algorithm and reimplement it as given in Assembly.
That is not the only problem nowadays, even processors within a given family may or may not have specific features (VT, for example) disabled. You'd think that there is a conspiracy going on...
It was probaby due to changes to Network Manager and how it detects 3G modem cards. YMMV. I have a Huawei E220 USB modem. It was fixed during Fedora 11, I use it everyday. However, I had to delete the old settings and recreate the modem entry. Glad to hear you got it working in F12.
My only issue with text files encrypted with gpg is that one of these days, I'd accidentally type the gpg decryption key on the command line instead of when prompted, and the key will show up nicely in the shell history. After the "Arrgh!!" has subsided I'd have to remember to purge the history file, etc.
Thanks for the info! I didn't particularly like xfig due to its interface; though you're right in that it has reasonably good integration with LyX.
I've been slowly trying to pick up LyX to replace MS Word for writing papers and technical documents in Comp. Sci.
Currently the biggest impediment in using LyX is the graphics / diagram integration. There is no easy way to edit the diagram from LyX (unlike MS Word where diagrams are embedded); it is a multi-step process where I'd edit the diagram in some other program (Dia is still not polished enough yet), export to EPS, then relink to the EPS in LyX if the filename were different.
The other problem is that EPS output is pretty spotty in most MS programs. Visio's EPS output sucks, to put it mildly, and often cause postscript errors. SVG export is still not well supported in ANY tool I've tried.
I'd be interested to know what kind of apps and formats works best when dealing with diagrams which needs to be cross-platform (Linux & Mac OS X).
I've never managed to get Word's autonumbering to work correctly in a continuously edited technical document despite many attempts to reset / modify the Heading X styles. The auto-formatting feature in MS Word is the worst culprit. This has always been MS Word's problem from the beginning (I'm currently using Office 2003 out of necessity).
I've never tried Office 2007, didn't want to invest any more energy on learning an incompatible interface on an application that does not deal well with technical documents.
LyX (LaTeX) is great for stuff like auto-numbering, but diagrams are a pain since there's no easy way to edit the diagram by double clicking on it, besides the lack of good diagramming tools that export to EPS format. Visio generates lousy EPS (deliberately?)
FrameMaker was the best of the lot. Unfortunately its market share and numerous bugs during the time Adobe bought it over meant that it's no longer a real contender for most people.
On the website URL, it states that it has been taken offline temporarily as of Jan 19, 2008 8PM CST. Sigh.
Just found it on the Seagate website. However, no md5 checksums provided for the ISO. Hope that the Flash utility has some internal checksumming...
I wish the ports system would have a better dependency resolution mechanism. Every time a new port comes out, it'd frequently trigger recompilations of various other dependencies whether it's needed or not when I specify a forced install. Ditto with removing older packages with multiple versions installed, it'd complain about bogus dependencies which do not affect the older version of the package. Whereas with RHEL (which I'm most familiar with), RPM would trigger a dependency package install only when it is actually needed.
The probably solution? TPM built into your PC. Then you can't complain about not being able to remove the DRM since it's built in. You PC would be just like a console. Not that it hasn't been tried before..... copy protection was pretty common in the 80's, then it was removed in the 90's due to hue and cry. I see it as a pendulum, we're just swinging the other way again right now, in a new DRM arms race.
Paradoxically, in British (English?) Medical Schools, a MD is a Bachelor's degree, and a postgraduate in Medicine earns the title of 'Mr.'
We're running native dual-stacked IPv6 in our lab at the university, and my pet peeve is that rouge routers announcing invalid routes will cause the IPv6 enabled systems (mostly Linux, in our case) to be misrouted (no route to host) until the interface is reset (ifdown/ifup). If there's a timeout mechanism associated with router advertisments, then this wouldn't be such a PITA issue. The other issue has to do with multihoming for BGP routing. AFAIK it's still a 'research problem'.
Prior to XP-SP2, it was possible to run WinXP Pro on my Fujitsu P2120 (256 MB -> 240 MB due to the x86 translator) at a reasonable speed. However, after upgrading to XP-SP2, it now takes almost 5 minutes from power on to the logged-in desktop after all the System Tray icons has stopped loading due to all the disk paging. Maybe the statement that it worked with 128 MB might have been true prior to XP-SP2, but it's definitely not true now.
I wonder if this meme was planted by spammers to help them bypass Spam filtering (and conditioning readers to intentionally mis-spelt text). I remembered that a lot of strangely spelt spam flooding my Inbox around the same time. Either that or Spammers latched on to this real quick.
I'm wondering what would be the option in future when IDE is totally depreciated. Any idea whether CF-to-SATA adapters exist?
>The syntax for "||" is:
>If command 1 fails, do command 2 - otherwise exit (where you used "command1 || command2").
But that's exactly what the Leopard Finder is doing!
Maybe he's the Leopard Finder's coder?
If your definition of blackhole routing is correct, then won't XP SP3 make life easier for botnets to launch DDos attacks, since the XP network stack would just drop the blackholed routes? I hope it's not a supposed 'fix' leading to an even greater problem.
> Allow multiple router announcements and behave gracefully. Use the prefix that best matches the destination.
I'm not sure how it could decide on the 'best' prefix to use if it is completely outside of the assigned subnet prefix. That should be the job of the router(s).
Actually what bugs me about IPv6 autoconfigured router prefixes is that it NEVER expires on the client machine (at least, not on Linux).
Someone bringing up a rogue radvd daemon would kill existing routes since it decides rather arbitrarily which router it would use as the default, and it usually ends up being the non-working rogue prefix.
The only workaround is to down the client interface and bring it up again to get a new autoconfigured prefix.