Based on the specs, this device only sends video from a computer (via a network) to the TV, but not from the TV to the computer. You'd think that if they were going to make a computer to TV/stereo interface that they would make it bidirectional to allow for TiVo-like functionality.
Oh, yea, how long until Apple comes out with one of these devices specially tailored for use with iLife (the iTunes/iPhoto/iMovie bundle). That would be killer.
The article states that they have "taped-out" the design. However, when I visited IBM-Austin last December (I gave a presentation on my research) they were still in the high-level idea phase. There is no way they could have decided on the design and completed it so quickly. My guess is this is a "test chip", like the one they did for Power4. Power4's test chip tested some of the critical circuits and such, but it was not the final design.
That said, it seemed like they were considering some pretty wild ideas. However, I remember hearing about plans for the Playstation 2 chip a couple of years before it shipped; at the time it was hard to fathom, but when it arrived it wasn't as big a leap as I thought it was going to be. (Though still quite impressive.)
I expect the Playstation 3 will be just as impressive, but not earth-shattering. They key will be how easy it is to write programs that take advantage of the raw computational power.
When viewing an HTML mail in Yahoo, it does the translation before it displays the mail for you. However, if you 'export' or download the message, it still looks fine. Thus, it looks as if the messages are not being changed when sent or received, they are only modified when being displayed through Yahoo's HTML webmail. Granted, based on the google searches, it is still causing lots of problems for users.
Looking at the technical specs, it appears that each drive has its own
controller. To quote from Apple's site:
"Each drive has an independent Ultra ATA/100 bus, an arrangement that
allows maximum individual drive performance without choking the
throughput of the other drives. SCSI is better than IDE when
controlling multiple disks. Apple must have done that math and
figured 4 IDE controllers and with IDE drives had a better
price/performance than a SCSI-based system."
I also don't think that IDE support hot swapping as well as SCSI.
However, It looks like having a controller per disk also allows Apple to get around
this (the new servers do support hot-swap drives). They probably just shut down the entire IDE controller for the
drive to allow hot swapping.
All in all, I think these new servers look very cool.
The ProScan VCR I have has an "auto-commercial-skip" feature like the one you describe. BTW, this isn't a feature for which I paid extra. My VCR isn't a high-end model or anything; I'm not even sure I was even aware it had this feature when I bought it.
After you record a show it rewinds the tape and scans what you recorded looking for commercial breaks. When you play back the tape, it begins fast forwarding at the beginning of a commercials and resumes normal play speed at the end of the commercials.
It works pretty well, though Law and Order sometimes confuses it with their scene transitions. However, since you watch the commercial go by, you can tell when it screws up and then you can just rewind or fast forward as needed.
If a mainstream VCRs I bought a couple of years ago can do this, why are so many corporations fixating on TiVOs?
Instead of offering xxx kbits/second and charging more per bit after a
certain usage threshold, the ISPs should sell a broadband connection
with a "peak" and "sustained" rating (e.g., 512kb peak and 56kb
sustained.) A users would receive bursts of 512kb throughput, but
after an hour or two straight at full throttle the ISP's router would
slowly limit throughput to the sustained rate.
One simple and well-known algorithm to implement this solution is a
token-bucket. (More information from Cisco's web site)
The basic idea is that you have a bucket that collects
token at some rate. This rate corresponds to the peak rate of
transfer. The bucket also has a maximum capacity which corresponds to
the size of the 'burst' you'll allow.
When a packet arrives and the bucket is non-empty, the packet is
forwarded and one token is removed from the bucket. When the bucket
is empty the packet is queued or dropped.
Going back to the above example, consider a token-bucket where tokens
arrive at 56kb/second, and the bucket can hold (60*60*512) kbits of
tokens. This bucket would allow full peak allows full use for a hour
or two, at which time the bucket would be close to empty and packets
could only be sent the sustained rate.
This kind of setup would not effect most users at all, but would limit
the worst offenders to 1/10th or 1/100th the bandwith usage.
A new device that will serve as a bridge between your home
TV/stereo and your Mac. It will do all the work of a non-portable MP3
player, DVD player, VCR, and maybe a gaming console.
It should connect to your TV and stereo via standard audio/video
input and output. It should connects to a Mac via firewire, Ethernet,
or airport (802.11). There should also be also a two-way remote
control for the device as well.
The device I'm proposing should allow a Mac to read and write
video/audio from a home TV/stereo. This allows the device to serve
MP3 audio from the Mac's harddrive, play quicktime video on the TV
(including iMovie created video), record and play back TV similar to
the TiVo, use it to play Internet-multiplayer games on your TV, and
display your digital pictures on the TV. I believe the system would
even allow the Mac to display a DVD from its drive on to the TV as
well. In addition, if you have a VCR, you can use the device to
import VHS home video into iMovie.
Now, let's look at the hardware required. All the box needs to
be able to do is encode and decode digitalaudio and maybe perform some
lightweight compression/de-compression. The box has an embedded
(slow & cheap) version of a PowerPC in it. The box has some amount of
DRAM, but no hard drive, since it can just use the Mac's hard drive.
It has just enough flash memory to net-boot from the connected Mac.
What operating system does the device run? Well, Darwin, of course!
Since the digital hub device boots from the Mac, it has almost no hard
state, and thus the software can be upgraded by installing new
software on the Mac.
The above description is actually quite similar to many of the
hardware firewall devices that use an embedding chip and Linux or
*BSD. You can buy these for under $100. Apple's new device would
needs to add a little bit of support for encoding and decoding, but I
would guess the final street price of the device would be under $200.
The biggest challenge with this device would providing enough
bandwidth to send TV quality video between the two systems. Apple's
wireless has 11 Mbs, or ~1MB/second bandwidth max. DVDs hold ~5 GBs
of data for ~2 hours of viewing. That's ~0.7MB/second for the
compressed stream. It should be pretty easy to downgrade the signal
from DVD quality to a TV quality signal, compress it a bit, and send
it from the Mac to the Hub staying well below the 1MB/second bandwidth
limit.
Clearly, this new device could take full advantage of Mac OS
X/Darwin. Darwin could be used on the device, and Mac OS X's
stability and multitasking support is needed to run the software for
it on the host Mac in the background. In fact, multiple processors
would work nicely for this application...
There have been rumors that Nintendo and Apple might be teaming
up. A device like this would be the perfect opportunity for Nintendo
to avoid fighting against the XBox and the PlayStation.
With a move like this, Apple could take control of the home entertainment
system.
(Yes, this is similar to the post I made on the previous topic today.)
Apple is going to announce a new "Digital Hub Device". Here is what I
think Apple should announce:
The new device will serve as a bridge between your home TV/stereo
and your Mac. It will do all the work of a non-portable MP3 player,
DVD player, VCR, and maybe a gaming console.
The digital hub should connect to your TV and stereo via standard
audio/video input and output. It connects to a Mac via firewire,
Ethernet, or airport (802.11). There should also be also a two-way
remote control for the device as well.
The new device should allow a Mac to read and write video/audio from
a home TV/stereo. This allows the device to serve MP3 audio from
the Mac's harddrive, play quicktime video on the TV (including
iMovie created video), record and play back TV similar to the TiVo,
use it to play Internet-multiplayer games on your TV, and display
your digital pictures on the TV. I believe the system would even
allow the Mac to display a DVD from its drive on to the TV as well.
In addition, if you have a VCR, you can use the device to import VHS
home video into iMovie.
Now, let's look at the hardware. All the box needs to be able to do
is encode and decode digitalaudio and maybe perform some
lightweight compression/de-compression. The box has an embedded
(slow & cheap) version of a PowerPC in it. The box has some amount
of DRAM, but no hard drive, since it can just use the Mac's hard
drive. It has just enough flash memory to net-boot from the
connected Mac. What operating system does the device run? Well,
Darwin, of course! Since the digital hub device boots from the Mac,
it has almost no hard state, and thus the software can be upgraded
by installing new software on the Mac.
The above description is actually quite similar to many of the
hardware firewall devices that use an embedding chip and Linux or
*BSD. You can buy these for under $100. Apple's new device would
needs to add a little bit of support for encoding and decoding, but
I would guess the final street price of the device would be under
$200.
The biggest challenge with this device would providing enough
bandwidth to send TV quality video between the two systems. Apple's
wireless has 11 Mbs, or ~1MB/second bandwidth max. DVDs hold ~5 GBs
of data for ~2 hours of viewing. That's ~0.7MB/second for the
compressed stream. It should be pretty easy to downgrade the signal
from DVD quality to a TV quality signal, compress it a bit, and send
it from the Mac to the Hub staying well below the 1MB/second
bandwidth limit.
Clearly, this new device will take full advantage of Mac OS
X/Darwin. Darwin will be running on the hub, and Mac OS X's
stability and multitasking support is needed to run the digital hub
in the background. In fact, multiple processors would work nicely
for this application...
There have been rumors that Nintendo and Apple might be teaming up.
This would be the perfect opportunity for Nintendo to avoid fighting
against the XBox and the PlayStation.
With this move, Apple could take control of the home entertainment
system. Maybe Apple will just release some new portable MP3 player,
but I don't think so...
Thanks to the DVR-7000's "Video Mode", DVD-R and DVD-RW discs can be played on home DVD players and DVD-ROM drives - the recording time is one hour in "V1 mode" and two hours in "V2 mode". Of course, Pioneer's new DVD recorder will be Macrovision-, CPRM*-, and CSS-protected.
* CPRM stands for "Copy Protection for Recordable Media". It refers to a copyright protection technology that records a "scramble" signal with the video content onto a recordable media that will not allow secondary copying of "once-only copy free" video contents.
Re:$ is made from HW, not SW
on
IBM Wants Linux
·
· Score: 1
"Everybody knows that the real money is made in hardware, not software".
IBM is all about services. They want to sell you a entire system (software & hardware), 24/7 service, and tech support -- an all in one 'solution'. The service part of things was about half of IBM's revenue last I checked.
FYI: High-end service and support was one of the main reasons that Compaq aquired Digital (DEC).
In the 90s, we'll probably see only ten real breakthroughs in computers.
Here are seven of them."
That 1989 NeXT poster goes on to say something like: Visit our showroom and you'll have a good idea where the remaining 3 breakthroughs will come from.
Remember, the first web browser was developed on the NeXT. If you go back and read the CACM article on the development of the web, you'll see some pretty NeXT screen shots.
NeXT had it right. They were just ten years ahead of their time.
CPU load balancing: If you're using many servers instead of
one, you need to make sure that your load is spread evenly across the
systems. For some applications, load imbalance can significantly
reduce efficiency. An SMP can dynamically balance between many tasks,
efficiently utilizing the processors.
Sharing other resources: SMPs allow the jobs to dynamically share
other (non-processor) resources such as RAM, swap, disk I/O bandwidth, etc.
For example, a web server can share a large file cache between all the
server processes, making efficient use of the DRAM.
One of my favorite uses of smaller SMPs is parallel compilation via
make -j. You can launch 2n jobs on n processors and get good speedup.
This does not work as well on a cluster setup. You can do it, it is
just a bit of a pain.
Of course this all depends on the cost premium of an SMP over
uniprocessors. This is more of the market effect of economies of scalar
rather than a fundamental issue with SMPs. If everyone bought SMPs,
they would have at worst a small cost premium. SMPs could even be
_cheaper_ (per processor) since you can share the DRAM, case, power supply, keyboard,
video card, and maybe even monitor (unless you run headless servers).
...and gnucash will automatically be installed and run. Oh, you have to be 'root' to do it, I'm sorry. Linux has this nice thing called 'permissions', so that you can't break anything unless you've logged in as root.
Why must everything in Linux be installed as root? Why can't RPMs and other packages be installed in my home directory and 'just work'? When you're root you can really screw things up for other users. I use Linux in a university environment where there is professional support staff. Thus, end users are strong discouraged from installing things as root. Trying to install new Linux software in my local user directory is almost impossible with some packages.
NEXTSTEP (now Apple's OS X) does a much better job with this. They have the concept of an application bundle (which is just a folder) that you can install in a system wide area (/LocalApps/) or in your user space (~/Apps/) and the application works the same either way.
It starts with transistors, talks about the
design of a basic computer, introduces assembly language, and ends
with an introduction to C. All of this in one or two semesters
designed to be taken as an intro. sequence or in parallel with a
more traditional track. I'd love to see a curriculum where the first
year includes a course that uses Java and another course that uses this
book.
Any new medium must offer something substantial for it to be adopted. In the case of CD's it was quality of music.
The music industry could decide to artificially sell this (or
another) new format for a lower cost than CDs. They could make some
bogus argument about the 'high cost' of CD pirating to justify the
high price of CDs. Lower prices could give people the (short term)
incentive to change format full of 'content protection.'
Yes, but if you don't artificially exlude [sic] clustered results
because you don't like what they say...
One of the well-known problems with TPC-C is that it uses a
hierarchical system where all the data is part of a particular
warehouse and only a small percent of transactions need to access any
cross-warehouse data. The little intra-warehouse communication
required is evenly distributed, so load imbalance is not a problem.
For this reason, the TPC-C benchmark can run more efficiently on a
cluster of servers than many real world OLTP setups. From what I've
been told, real clustered DBMS setups suffer from load imbalance and
are even more difficult to setup and tune than 'single instance'
systems. The majority of OLTP systems in the field don't use clusters
for performance.
That said, one of the best uses of small clusters is to provide fault
tolerance and high availability of data, but that is a different setup
than the 'clustered' TPC-C results you mentioned.
Right now the top
non-clustered TPC-C score is held IBM's s80 system. TPC-C (not SPEC) is considered by
many to be the most important server benchmark.
The system from the benchmark report has 24 RS64-IV 64-bit processors
running at 600 Mhz with 96GB (yes, GB) of system DRAM. Each processor
has 128kB L1 data cache, 128kB L1 instruction cache, and a 16MB L2
cache. The chips also support course-grain multithreading (simpler,
but similar to SMT).
(600 Mhz sounds slow until you realize that it uses a simple, very
efficient 5-stage pipeline. Intel and others achieve high clock rates
through deep piplines and rely on branch prediction and other
techniques to keep the pipe full. Branch mispredictions and cache
misses can kill the actual performance of these chips on real server
code.)
This system with 24 processors outperforms HP's 48 processor
"SuperDome" and Sun's 64 processor EU10k (though the UE10k is an old
system by now, it is the fastest server Sun is shipping.)
The above system is not using the Power3 chip from the posted story.
You can bet IBM will port Linux to this beast next. We won't see a 24 processor systems with Linux
right away, but an s80-like system would make a sweet 4-processor
Linux server.
One last note: these systems are not vapor-ware. A 12-processor system with an earlier version of the same processor has been shipping since the summer of '98.
The real reason you want 64-bits is for memory addressing. 64-bit means having 64-bit general purpose registers, a page table in the OS that can support the new 64-bit virtual address space, etc. Having a 64-bit bus is not really the issue.
That said, it would be easier to go PowerPC 64-bit, but when I went to find out information from Motorola's site, all I could find was information about embeded and low-power, low-performance PowerPC. There was a roadmap that listed a high-performace "G5" chip with some 64-bit mode, but zero details.
I don't think Motorola is focused enough on high performance chips. Macs that use these chips will fall farther and farther behind Intel/AMD performance.
Apple should wait for AMD's x86-64 before moving OSX to x86. The memory (virtual and physical) addressing limitations of 32-bit PowerPCs are going to force able to Apple to start the move to 64-bits in the next couple of years. They either have to go PowerPC 64-bit, or they could jump to x86-64.
BTW, does anyone have any idea what PowerPC chip (for the Mac) follows the G4?
Not to mention that AMD already is licensing the most important thing it could get from Transmeta -- the code-morphing tech so that they can simulate their upcoming chips
I think AMD has a different plan for simulating ClawHammer than using
Transmeta technology. AMD recently announced it is teaming up with Virutech a Swedish startup that
produces Simics, a full
system simulator. You can find out more see AMD's press
release and Virtutech's
press release.
What then? By producing drugs that alleviate the symptoms but which do
not cure the disease, making the patients dependent on those drugs for
the rest of their life.
Drug companies, like all companies, are looking to make money. If
drug companies are only making drugs to alleviate symptoms, maybe the
government should step in and provide incentives
to encourage these companies to develop drugs that actual cure or
prevent disease.
It is not difficult to think of drugs, such as antibiotics and
vaccines, that are counter-examples to your claim that drug companies
focus on producing drugs that only alleviate symptoms.
That said, my original point still stands. Drug companies are not
going to produce drugs (even ones that only alleviate symptoms)
without some intellectual property/patent protection.
Individuals and corporations need some sort of incentive to perform research and, for example, produce
new drugs. A basc motivation for
patent protection is to grant some tangible 'value' to these
pursuits.
Companies perform a cost/benefit analysis trying to decide if
they should research, produce, seek FDA approval, and then mass
produce a new medicine. With reasonable patent protection (not
unreasonable like Amazon's 1-click) the drug company can spend more
money on R&D with some hope to recover this investment before other
companies can benefit from their hard work.
If we don't give the drug company financial incentives to produce new
drugs through patent protection, they will slow their rate of researching new drugs. Is that in
society's best interest?
We should reform patent laws, but I think we still want them.
Slashdot Mirror
Based on the specs, this device only sends video from a computer (via a network) to the TV, but not from the TV to the computer. You'd think that if they were going to make a computer to TV/stereo interface that they would make it bidirectional to allow for TiVo-like functionality.
Oh, yea, how long until Apple comes out with one of these devices specially tailored for use with iLife (the iTunes/iPhoto/iMovie bundle). That would be killer.
Seymour Cray said it best: "If you were plowing a field, which would you rather use? Two strong oxen or 1024 chickens?"
That said, it seemed like they were considering some pretty wild ideas. However, I remember hearing about plans for the Playstation 2 chip a couple of years before it shipped; at the time it was hard to fathom, but when it arrived it wasn't as big a leap as I thought it was going to be. (Though still quite impressive.)
I expect the Playstation 3 will be just as impressive, but not earth-shattering. They key will be how easy it is to write programs that take advantage of the raw computational power.
When viewing an HTML mail in Yahoo, it does the translation before it displays the mail for you. However, if you 'export' or download the message, it still looks fine. Thus, it looks as if the messages are not being changed when sent or received, they are only modified when being displayed through Yahoo's HTML webmail. Granted, based on the google searches, it is still causing lots of problems for users.
"Each drive has an independent Ultra ATA/100 bus, an arrangement that allows maximum individual drive performance without choking the throughput of the other drives. SCSI is better than IDE when controlling multiple disks. Apple must have done that math and figured 4 IDE controllers and with IDE drives had a better price/performance than a SCSI-based system."
I also don't think that IDE support hot swapping as well as SCSI. However, It looks like having a controller per disk also allows Apple to get around this (the new servers do support hot-swap drives). They probably just shut down the entire IDE controller for the drive to allow hot swapping.
All in all, I think these new servers look very cool.
After you record a show it rewinds the tape and scans what you recorded looking for commercial breaks. When you play back the tape, it begins fast forwarding at the beginning of a commercials and resumes normal play speed at the end of the commercials.
It works pretty well, though Law and Order sometimes confuses it with their scene transitions. However, since you watch the commercial go by, you can tell when it screws up and then you can just rewind or fast forward as needed.
If a mainstream VCRs I bought a couple of years ago can do this, why are so many corporations fixating on TiVOs?
One simple and well-known algorithm to implement this solution is a token-bucket. (More information from Cisco's web site) The basic idea is that you have a bucket that collects token at some rate. This rate corresponds to the peak rate of transfer. The bucket also has a maximum capacity which corresponds to the size of the 'burst' you'll allow. When a packet arrives and the bucket is non-empty, the packet is forwarded and one token is removed from the bucket. When the bucket is empty the packet is queued or dropped.
Going back to the above example, consider a token-bucket where tokens arrive at 56kb/second, and the bucket can hold (60*60*512) kbits of tokens. This bucket would allow full peak allows full use for a hour or two, at which time the bucket would be close to empty and packets could only be sent the sustained rate.
This kind of setup would not effect most users at all, but would limit the worst offenders to 1/10th or 1/100th the bandwith usage.
About time we had real ads, not these wimpy little banner ads and stuff. Go Slashdot!
Oh, uh, April 1st? Nevermind.
The original Microprocessor Forum presentation is available:l oadableAssets/MPF_Hammer_Presentation.PDF
http://www.amd.com/us-en/assets/content_type/Down
- A new device that will serve as a bridge between your home
TV/stereo and your Mac. It will do all the work of a non-portable MP3
player, DVD player, VCR, and maybe a gaming console.
- It should connect to your TV and stereo via standard audio/video
input and output. It should connects to a Mac via firewire, Ethernet,
or airport (802.11). There should also be also a two-way remote
control for the device as well.
- The device I'm proposing should allow a Mac to read and write
video/audio from a home TV/stereo. This allows the device to serve
MP3 audio from the Mac's harddrive, play quicktime video on the TV
(including iMovie created video), record and play back TV similar to
the TiVo, use it to play Internet-multiplayer games on your TV, and
display your digital pictures on the TV. I believe the system would
even allow the Mac to display a DVD from its drive on to the TV as
well. In addition, if you have a VCR, you can use the device to
import VHS home video into iMovie.
- Now, let's look at the hardware required. All the box needs to
be able to do is encode and decode digitalaudio and maybe perform some
lightweight compression/de-compression. The box has an embedded
(slow & cheap) version of a PowerPC in it. The box has some amount of
DRAM, but no hard drive, since it can just use the Mac's hard drive.
It has just enough flash memory to net-boot from the connected Mac.
What operating system does the device run? Well, Darwin, of course!
Since the digital hub device boots from the Mac, it has almost no hard
state, and thus the software can be upgraded by installing new
software on the Mac.
- The above description is actually quite similar to many of the
hardware firewall devices that use an embedding chip and Linux or
*BSD. You can buy these for under $100. Apple's new device would
needs to add a little bit of support for encoding and decoding, but I
would guess the final street price of the device would be under $200.
- The biggest challenge with this device would providing enough
bandwidth to send TV quality video between the two systems. Apple's
wireless has 11 Mbs, or ~1MB/second bandwidth max. DVDs hold ~5 GBs
of data for ~2 hours of viewing. That's ~0.7MB/second for the
compressed stream. It should be pretty easy to downgrade the signal
from DVD quality to a TV quality signal, compress it a bit, and send
it from the Mac to the Hub staying well below the 1MB/second bandwidth
limit.
- Clearly, this new device could take full advantage of Mac OS
X/Darwin. Darwin could be used on the device, and Mac OS X's
stability and multitasking support is needed to run the software for
it on the host Mac in the background. In fact, multiple processors
would work nicely for this application...
- There have been rumors that Nintendo and Apple might be teaming
up. A device like this would be the perfect opportunity for Nintendo
to avoid fighting against the XBox and the PlayStation.
With a move like this, Apple could take control of the home entertainment system. (Yes, this is similar to the post I made on the previous topic today.)- The new device will serve as a bridge between your home TV/stereo
and your Mac. It will do all the work of a non-portable MP3 player,
DVD player, VCR, and maybe a gaming console.
- The digital hub should connect to your TV and stereo via standard
audio/video input and output. It connects to a Mac via firewire,
Ethernet, or airport (802.11). There should also be also a two-way
remote control for the device as well.
- The new device should allow a Mac to read and write video/audio from
a home TV/stereo. This allows the device to serve MP3 audio from
the Mac's harddrive, play quicktime video on the TV (including
iMovie created video), record and play back TV similar to the TiVo,
use it to play Internet-multiplayer games on your TV, and display
your digital pictures on the TV. I believe the system would even
allow the Mac to display a DVD from its drive on to the TV as well.
In addition, if you have a VCR, you can use the device to import VHS
home video into iMovie.
- Now, let's look at the hardware. All the box needs to be able to do
is encode and decode digitalaudio and maybe perform some
lightweight compression/de-compression. The box has an embedded
(slow & cheap) version of a PowerPC in it. The box has some amount
of DRAM, but no hard drive, since it can just use the Mac's hard
drive. It has just enough flash memory to net-boot from the
connected Mac. What operating system does the device run? Well,
Darwin, of course! Since the digital hub device boots from the Mac,
it has almost no hard state, and thus the software can be upgraded
by installing new software on the Mac.
- The above description is actually quite similar to many of the
hardware firewall devices that use an embedding chip and Linux or
*BSD. You can buy these for under $100. Apple's new device would
needs to add a little bit of support for encoding and decoding, but
I would guess the final street price of the device would be under
$200.
- The biggest challenge with this device would providing enough
bandwidth to send TV quality video between the two systems. Apple's
wireless has 11 Mbs, or ~1MB/second bandwidth max. DVDs hold ~5 GBs
of data for ~2 hours of viewing. That's ~0.7MB/second for the
compressed stream. It should be pretty easy to downgrade the signal
from DVD quality to a TV quality signal, compress it a bit, and send
it from the Mac to the Hub staying well below the 1MB/second
bandwidth limit.
- Clearly, this new device will take full advantage of Mac OS
X/Darwin. Darwin will be running on the hub, and Mac OS X's
stability and multitasking support is needed to run the digital hub
in the background. In fact, multiple processors would work nicely
for this application...
- There have been rumors that Nintendo and Apple might be teaming up.
This would be the perfect opportunity for Nintendo to avoid fighting
against the XBox and the PlayStation.
With this move, Apple could take control of the home entertainment system. Maybe Apple will just release some new portable MP3 player, but I don't think so...From the web site:
"Everybody knows that the real money is made in hardware, not software".
IBM is all about services. They want to sell you a entire system (software & hardware), 24/7 service, and tech support -- an all in one 'solution'. The service part of things was about half of IBM's revenue last I checked.
FYI: High-end service and support was one of the main reasons that Compaq aquired Digital (DEC).
That 1989 NeXT poster goes on to say something like: Visit our showroom and you'll have a good idea where the remaining 3 breakthroughs will come from.
Remember, the first web browser was developed on the NeXT. If you go back and read the CACM article on the development of the web, you'll see some pretty NeXT screen shots.
NeXT had it right. They were just ten years ahead of their time.
- CPU load balancing: If you're using many servers instead of
one, you need to make sure that your load is spread evenly across the
systems. For some applications, load imbalance can significantly
reduce efficiency. An SMP can dynamically balance between many tasks,
efficiently utilizing the processors.
- Sharing other resources: SMPs allow the jobs to dynamically share
other (non-processor) resources such as RAM, swap, disk I/O bandwidth, etc.
For example, a web server can share a large file cache between all the
server processes, making efficient use of the DRAM.
One of my favorite uses of smaller SMPs is parallel compilation via make -j. You can launch 2n jobs on n processors and get good speedup. This does not work as well on a cluster setup. You can do it, it is just a bit of a pain.Of course this all depends on the cost premium of an SMP over uniprocessors. This is more of the market effect of economies of scalar rather than a fundamental issue with SMPs. If everyone bought SMPs, they would have at worst a small cost premium. SMPs could even be _cheaper_ (per processor) since you can share the DRAM, case, power supply, keyboard, video card, and maybe even monitor (unless you run headless servers).
Why must everything in Linux be installed as root? Why can't RPMs and other packages be installed in my home directory and 'just work'? When you're root you can really screw things up for other users. I use Linux in a university environment where there is professional support staff. Thus, end users are strong discouraged from installing things as root. Trying to install new Linux software in my local user directory is almost impossible with some packages.
NEXTSTEP (now Apple's OS X) does a much better job with this. They have the concept of an application bundle (which is just a folder) that you can install in a system wide area (/LocalApps/) or in your user space (~/Apps/) and the application works the same either way.
It starts with transistors, talks about the design of a basic computer, introduces assembly language, and ends with an introduction to C. All of this in one or two semesters designed to be taken as an intro. sequence or in parallel with a more traditional track. I'd love to see a curriculum where the first year includes a course that uses Java and another course that uses this book.
The music industry could decide to artificially sell this (or another) new format for a lower cost than CDs. They could make some bogus argument about the 'high cost' of CD pirating to justify the high price of CDs. Lower prices could give people the (short term) incentive to change format full of 'content protection.'
One of the well-known problems with TPC-C is that it uses a hierarchical system where all the data is part of a particular warehouse and only a small percent of transactions need to access any cross-warehouse data. The little intra-warehouse communication required is evenly distributed, so load imbalance is not a problem.
For this reason, the TPC-C benchmark can run more efficiently on a cluster of servers than many real world OLTP setups. From what I've been told, real clustered DBMS setups suffer from load imbalance and are even more difficult to setup and tune than 'single instance' systems. The majority of OLTP systems in the field don't use clusters for performance.
That said, one of the best uses of small clusters is to provide fault tolerance and high availability of data, but that is a different setup than the 'clustered' TPC-C results you mentioned.
The system from the benchmark report has 24 RS64-IV 64-bit processors running at 600 Mhz with 96GB (yes, GB) of system DRAM. Each processor has 128kB L1 data cache, 128kB L1 instruction cache, and a 16MB L2 cache. The chips also support course-grain multithreading (simpler, but similar to SMT).
(600 Mhz sounds slow until you realize that it uses a simple, very efficient 5-stage pipeline. Intel and others achieve high clock rates through deep piplines and rely on branch prediction and other techniques to keep the pipe full. Branch mispredictions and cache misses can kill the actual performance of these chips on real server code.)
This system with 24 processors outperforms HP's 48 processor "SuperDome" and Sun's 64 processor EU10k (though the UE10k is an old system by now, it is the fastest server Sun is shipping.)
The above system is not using the Power3 chip from the posted story. You can bet IBM will port Linux to this beast next. We won't see a 24 processor systems with Linux right away, but an s80-like system would make a sweet 4-processor Linux server.
One last note: these systems are not vapor-ware. A 12-processor system with an earlier version of the same processor has been shipping since the summer of '98.
That said, it would be easier to go PowerPC 64-bit, but when I went to find out information from Motorola's site, all I could find was information about embeded and low-power, low-performance PowerPC. There was a roadmap that listed a high-performace "G5" chip with some 64-bit mode, but zero details.
I don't think Motorola is focused enough on high performance chips. Macs that use these chips will fall farther and farther behind Intel/AMD performance.
Apple should wait for AMD's x86-64 before moving OSX to x86. The memory (virtual and physical) addressing limitations of 32-bit PowerPCs are going to force able to Apple to start the move to 64-bits in the next couple of years. They either have to go PowerPC 64-bit, or they could jump to x86-64. BTW, does anyone have any idea what PowerPC chip (for the Mac) follows the G4?
I think AMD has a different plan for simulating ClawHammer than using Transmeta technology. AMD recently announced it is teaming up with Virutech a Swedish startup that produces Simics, a full system simulator. You can find out more see AMD's press release and Virtutech's press release.
Drug companies, like all companies, are looking to make money. If drug companies are only making drugs to alleviate symptoms, maybe the government should step in and provide incentives to encourage these companies to develop drugs that actual cure or prevent disease.
It is not difficult to think of drugs, such as antibiotics and vaccines, that are counter-examples to your claim that drug companies focus on producing drugs that only alleviate symptoms.
That said, my original point still stands. Drug companies are not going to produce drugs (even ones that only alleviate symptoms) without some intellectual property/patent protection.
Companies perform a cost/benefit analysis trying to decide if they should research, produce, seek FDA approval, and then mass produce a new medicine. With reasonable patent protection (not unreasonable like Amazon's 1-click) the drug company can spend more money on R&D with some hope to recover this investment before other companies can benefit from their hard work.
If we don't give the drug company financial incentives to produce new drugs through patent protection, they will slow their rate of researching new drugs. Is that in society's best interest?
We should reform patent laws, but I think we still want them.