Having read the articles that were easy to get to, and the abstract of the PhD student: this is buzzword bollocks. There is no innovation in what they have done.
Me quoting you:
...this is buzzword bollocks.
Seems a pretty accurate sense of what you said.
Again, you:
Interesting research, as it's always good to see people explore different designs, but it sounds overhyped and I believe that it has zero commercial appeal.
And me quoting you:
I believe that it has zero commercial appeal.
Again, your original sentence concentrated on the perceived commercial value of the project, not the fact that it's a research project. You intended to shift the debate from one of exploring the design space in a relatively unfettered manner to whether the project serves some narrow economic interest. You did the same when comparing TRIPS to existing commercial architectures when running media applications, ignoring the actual future problems that TRIPS is designed to overcome. I simply focused on the argument you actually made.
Are you now unwilling to stand behind that argument? There's nothing wrong with changing your argument but there is something wrong with doing so while trying to obfuscate the change in debate with attacks.
From your other posts I would guess that you are (loosely) associated with the project / people on it.
You would guess wrong.
Please back up your claim about breaking code up into packets (preferably with a paper citation), because if they have done that I would like to read it.
Perhaps you would care to examine the TRIPS publications page where there are several papers on the compiler available.
As someone who did his PhD in an area that has no (foreseeable) commercial application I don't really need an explanation of this from you.
Nod to you for referencing transport-triggered architectures and TRIPS' relationship to them. Good job!
But I must disagree with you about scheduling. Scheduling is HARD. It's hard because there are lots of unknowns like memory latency and how the OOO engine will actually execute the code. The TRIPS compiler doesn't so much schedule the operations themselves as it does the communication. It has to worry about things like routing distance and so on. Yes, a naive implementation would be straightforward but doing something intelligent is often much more difficult.
Thanks for the reference to your work. I'll definitely take a look.
No, it isn't. The TRIPS group has done some really interesting things with compilers, for example. They've managed to have the compiler break up code into packets and schedule them on the processor array so that dependencies flow nicely across the grid. That is not an easy problem to tackle. This is very good research.
I believe that it has zero commercial appeal.
That's not the point of research. The point of research is to explore problems no one has tackled before, of course always with an eye toward future technology trends.
Which brings up a good question: what are they counting as ILP? It's been a while since I read the papers so I'll have to go and double-check, but if they're counting predicated instructions that get squashed in their "instructions simultaneously executing" number, it's not a fair comparison to current processors which report IPC as actually instructions that do work.
That said, IIRC TRIPS uses some interesting cache technology to help with speculation and relies on the compiler much more than a standard OOO architecture. If you're running the right kind of codes, there's a rather large amount of parallelism out there to exploit. Even if you're NOT running the right kinds of codes, you can still get a surprising amount of parallelism by removing the stack pointer if your branch prediction (or predication) and latency tolerance is good enough. I don't recall whether TRIPS makes use of this available parallelism or not.
Doug was one of the students who worked on the MultiScalar project at Wisconsin over a decade ago. Lots of good research out of that project and TRIPS bears more than a passing resemblance to MultiScalar. But, of course, it's different. The compilation model is much simpler and in the interim between the projects there have been a number of advances in compiler/architecture cooperation.
TRIPS is definitely not a vector processor. As another poster said, it's close to a dataflow machine, but it also has some of the characteristics of fine-grained multithreading. I'd place it in the spectrum as an architecture engineered to be as close to dataflow as possible while still being practical.
Back when I read the papers, when they had simulator results, it struck me that the individual scalar processors were pretty aggressive, very out-of-order, etc. Complex and difficult to design. I am very curious about what they were able to do in actual silicon. Looks like it's time to read some more papers.:)
Come on now, you're blaming an entire belief system for this tragedy?
This guy did nothing recognizable as Christian, regardless of what he said. Yes, there are despicable people who call themselves "Christian." In fact scripture itself tells us to expect them. Although, let's be VERY clear, it is not a "future telling" set of documents!
Judeo-Christian prophecy is quite different and apart from mere "fortune telling." It is at the heart of the faith tradition, a challenging of society to be something better than it is. There are many, many, many, MANY Christians and people of other faiths doing good work, prophetic work, for real justice in the world. Ghandi was a prophet. Dr. King was a prophet.
I can't speak for other religions (though I suspect their followers would agree with me), but Christian faith is not at all guilt-ridden and does not promote self-loathing. Yes, many people teach it that way and that's tragic. But a good faith community will lift one up, not put one down. I've been in good communities and bad ones. I tend to stay with the good ones.
The core teaching of Christianity is quite simple: God cares about human beings, who are inherently good. It is our responsibility to use our gifts and talents to make the world a better place for everyone.
Real ID was passed back as part of an "essential" Iraq funding package. As such it had no separate committee hearings. And at the time, guess who controlled Congress?
Many states are opposed to this not due to privacy concerns but simply because it's another unfunded federal mandate. Minnesota estimates it will cost the state $31 million over five years. Total national costs have ballooned to something like $17 billion. Congress allocated all of $40 million to pay for it.
I know it would take some work for me to produce a certified birth certificate or passport to get a license.
It will be interesting to see what happens when boarding restrictions, etc. go into effect. If this law isn't killed outright at the federal level, I believe it will be effectively ignored in many situations.
Minneapolis E-Democracy Issue Forum - Not a blog or web forum, a mailing list. Mailing lists are way better than web forums. Lots of good local information on events, politics, etc.
Polinaut - When the mood strikes me and/or I want to see if anything I did up at the Minnesota Capitol made the news.
The Strib - Only for the netlets (letters to the editor posted online only). For everything else it's strictly reading the hardcopy for me. It's too hard on the eyes to read from a screen.
It's not that much of a pig. As a compiler target, it's actually quite nice, except for the occasional register constraints (shift in *CX, etc.). Squeezing performance out can be a pain because of TIMTOWTDI. That makes instruction selection, register allocation and scheduling interact in weird and fascinating ways.
The ModR/M encoding is responsible for the limited register set, but at the time the ISA was created, it was a reasonable tradeoff. X86-64 fixed some of those problems at the expense of more prefixes.
The main problem with the ISA is hardware decoding. Icaches are a pain because you've got to find instruction boundaries. I can imagine that prefixes make the combinational logic rather interesting. Probably a lot of power goes into the decode part of the chip.
But from a programmer-level view, the ISA is actually fairly nice.
The only reason suburbs appear to be less expensive places to live than cities is because we hide the cost of living there. We build the freeways, sewers and other utilities needed for suburban homes but don't reflect that cost in home prices or property taxes. We subsidize the oil going into our cars. The Federal Housing Administration gave grants for suburban construction post-WWII but not for redeveloping cities. Mass transit support has been decimated, making it even harder for those in the cities to get to jobs.
When you learn the history of Detroit, you find its story repeated all over the country. Very prosperous minority neighborhoods were plowed over by the interstate system, destroying an entire economy in the process. We lost untold cultural treasures simply because our racism made it easy to destroy the homes and businesses of a large portion of our community.
Once the tax base is removed from the city, services decrease, poverty concentrates and crime tends to increase. Why does crime increase? Poor schools give young people no hope. The lack of tax base means that urban schools are at a severe disadvantage to suburban schools. The lack of jobs means that young people find other ways to occupy their time.
Fear, ignorance and prejudice causes concentrated poverty because forces manipulate those fears to divide us. Realtors steer whites to white communities, banks redline parts of town, making it more difficult to obtain loans in certain areas.
But even so, people are choosing to live in cities again. Sprawl has gone so far overboard that even with all of the cost hiding, people are spending more money on transportation from the exurbs than they do on their mortgages. I believe we're going to see a radical shift in our living patterns over the next 50 years and we had better begin investing in our infrastructure (e.g. mass transit) to make the transition as smooth as possible.
But then, something called the "ghetto" arose, driving people to the suburbs so they could keep some of the advantages of the city while not having to sleep in their bathtubs at night.
You've got it exactly backward. Cities declined because people moved out. All sorts of policies were put into place that drove white flight: Federal Housing Administration policy that rewarded new suburban construction over rehabilitation, redlining by lenders, steering by realtors, etc. Some of this stuff still goes on today. Suburbia arose due to deliberate government policy. Robert Moses was the instigator. It wasn't an accident.
Watch this fantastic documentary (parts I & II) about Detroit to learn more. It is not at all the case that cities necessarily decline into wastelands. In many parts of Europe, for example, the concentrated poverty is in suburbia (remember the Paris suburb riots?), which is of course its own problem. Concentrated poverty is the real issue, not cities vs. suburbs. It's our fear, ignorance and prejudice that causes it.
madhatter is right. Intel made power consumption a key part of their strategy and it is absolutely the correct one. Even in the server market, power is a huge deal. The power to run a datacenter is becoming an increasingly large part of the overall cost of providing services. Heck, even in the HPC market, we're seeing customers demand flops/watt.
AMD had better address this soon or they're going to get eaten alive. Right now, Hypertransport is what's keeping them in the game. When CSI comes along, AMD had better have an answer.
Even some cities like Atlanta are too spread out to have an efficient transit system.
That's a common misperception. It's a lie put forth by the anti-transit crowd to stop investment. The truth is that Atlanta is putting big money into transit, as is Dallas, Denver, Los Angeles, Phoenix and a whole lot of other places that one might consider highly sprawled.
The fact is that practically everywhere there has been major investment in transit, there has been major increase in transit usage. People use transit when it is convenient. That means there needs to be enough service for people to get to work, to the doctor, to the grocery and so on and be certain they can get back without waiting half an hour for a transfer.
Here in the Twin Cities, transit ridership is at a 22-year high, despite 6 years of fare increases and service cuts over the last 7 years. Our bus and rail system is pathetic, but usage is still rising. There is huge demand for transit here and ordinary people (i.e. not lobbyists) are practically knocking down doors at the State Capitol to get funding increased.
Intercity rail is picking up steam in the U.S. For example, the Midwest Regional Rail Initiative would build a network of high-speed rail connecting major cities in the midwest. Amtrak ridership is rising, along with Amtrak revenue.
Transit is a big part of how we will handle the energy and environmental crises we face. Yes the Bush administration proposes to cut $350 million annually from the Federal Transit Administration (FTA). The FTA itself is well-known for how it uses its rulemaking to kill rail projects across the country. I was recently in D.C. talking to our Minnesota delegation about this. Chairman Oberstar will hold oversight hearings and look into this.
We're seeing the start of a major shift in our transportation patterns here in the U.S. and it's not just in the big cities. Medium-sized cities and even rural areas are demanding major investment in transit. It's a good thing, too, because the single-occupancy vehicle has run out of gas. The car will and should always be with us but we can no longer sustain the ever-expanding sprawl and freways to support it without investing in alternative modes.
Language is important. It's how we express ourselves. If one is
not willing to take care in writing and speaking, he or she may not
give the best impression -- and impressions matter.
There's a difference between error-free writing and proper editing
to fix mistakes. No one writes error-free. Editors exist to catch
these types of problems.
But to have the attitude going into the process that spelling and
grammar don't matter is to isolate oneself in a small group of people
that is taken less seriously by the rest of the world.
You've all got it wrong. REAL ID isn't about terrorism, it's about immigration.
The law exists for one reason: some people want to see other people denied basic human rights and opportunity. While the law has scary slippery-slope scenarios, the real danger is what it will do the minute it goes into action: brand some people as lesser than others.
Amen! It's amazing to me the number of people who don't understand this. TIMTOWTDI is the worst aspect of Perl. Well, maybe not the very worst. I'd give that to the OO extensions (and I'm a long-time C++ fan).
TIMTOWTDI pretty much guarantees that no one else will be able to understand your code. It's great for being "clever" and for stroking your ego by impressing your friends with your one-liners, but that won't fly in software that has to be maintained.
Perl is fantastic for what it was originally intended to do: process text. Perl's use of regular expressions is right on. Perl was never meant to be used as a database, for large applications or anything beyond quick & dirty tools.
I use Perl a lot to process reports and create tables for spreadsheets and graphing programs. It works great for that kind of thing. But I would never use it in a large and complex project.
It's fairly easy to tell the differences between two performances of the same piece. Is the pedaling the same? How about dynamics? Tempo? How is rubato used? Are the accents played similarly?
I'm by no means an expert on classical music (traditional jazz is more my thing) but even I can tell the difference between someone who is technically proficient but lacks interpretation and a real pro who understands how the whole piece fits together.
My guess is that the copied performances were not all that well known. Even if a performer is well-known, it's not necessarily the case that all of his or her performances are of such quality that they are universally recognized.
Thank you! I switched to Lilypond several months ago and never looked back. It is so much more flexible than Finale and its ilk due to the fact that it isn't constrained by a graphical representation. I also find that writing music in text is a lot faster than point-and-click or even recording and going back to adjust all of the quantizing problems.
I love the ability to use music variables to hold repeating sequences. I love the programmability (even better with the new streams model). It's extremely easy to write parts for each instrument and mix-and-match them into different scores. I find that, for example, some people in choirs like to see the full SATB parts in a traditional two-staff layout, others prefer a four-staff layout while some prefer just to see their own part. The pianist really wants to see the SATB put on a standard two-staff piano score. No problem with lilypond, I can tailor the presentation to each individual choir member if I wish.
And it makes beautiful engravings, too!
In my opinion, Lilypond completely outclasses commercial and proprietary music scoring software.
Immersive gaming (CAVE in the home, advanced AI, physics & video, etc.)
I think things get particularly interesting thinking about many-core in handheld devices. Software radio could be tremendously useful there. Route planning using a network of GPS-enabled handheld devices would be cool. Background searching could be used for a lot of things.
100 cores is not massively parallel. The kind of scaling we're talking
about is much higher. Think thousands of cores each with hundreds of threads.
This is the kind of scaling that weather centers are just starting to reach today. It's the kind of scaling that will require a radical rethinking of how consumer software is designed and what tools we need to make that design process easier.
In this world, software is king. You won't care who your chip vendor is. You'll care who provides your compiler, debugger, performance analysis tools and other such things.
No, it isn't. It's classic Grover Norquist anti-tax, anti-government rhetoric packaged up in populist clothing. We levy taxes for a very important reason: they fund the services necessary for us to maintain our society and prosper.
It's not insightful to blather on about government inefficiency. You've got to show where it can be improved. More often than not, government programs are running on barebones budgets. There's no fat to cut.
I'm sick of right-wingers fanning the flames of fear and isolation by claiming that there's not enough to go around and that we each have to fight to get our own. I'm sick of those who say we're too poor to meet the needs of everyone in this country. We're the richest country in the history of the world, for God's sake!
Taxes on internet and mail-order purchases are long overdue. I always try to buy locally so that my city, county and state can get just a little more of the funding they need to operate. But I know that not everyone is thinking about that. I've worked up at the Minnesota state legislature (on my own dime and my own time), I've talked with legislators and I've had discussions with staff. I know about the needs of the people in our state and how they are not being met because there are certain elements that want to perpetuate the myth of scarcity for their own benefit.
It's incredibly selfish to whine about a small tax increase when there are homeless children going without food, when low-income people can't get to a job because there's no bus and when funding for neighborhood libraries is being slashed. Doubly so when you complain about having to pay that extra $20 for your shiny new flat-panel display.
The sad thing is, most people buy into the rhetoric without ever understanding or trying to understand what the real situation is. People are aware that something is not right with the way our country is headed. When I point out that the nonsense anti-government and, ultimately, anti-community fear-laced rhetoric is at the root of the problem, they begin to understand and support raising new taxes for services they want.
We make choices. So far, we've chosen to leave people behind in a big way. It's an immoral choice. But it's the choice we've made. We can eradicate poverty and provide opportunity for all. We just have to choose to do it. That means choosing to raise revenue.
Governor Pawlenty, I make a good salary. Tax me a little more so that those less fortunate can have a chance at the same life that I enjoy. Please.
She's not talking about Word. She's talking about ISV codes. Like NASTRAN. Her point is that these codes are not written to scale to thousands of cores. And she's right.
GPGPU and video games are not the same as bioinformatics or particle physics. These are heavy-duty codes, some of which have a very hard time scaling to large numbers of processors. Custom codes are usually much better about this than ISV codes.
Her point is that to use these new architectures, the software stack has to get much better. She's spot on. Certainly PGI and PathScale have not shown yet that they can keep up. I believe they will get there eventually, but it may be much later than we'd like. We will see.
But in my experience (and I'm highly curious of others reading this) some of the best songs in my and my band's repertoirs are ones that just "came out", from either screwing around and stumbling onto things that rock, to making fun of something that happened, etc.
Right on. Deadlines really help too. Last fall we were planning a public meeting for the community organizing group I work with. I told the planning team I was going to compose a choir piece for the event. I had a month to do it and had never written anything for a full choir before.
I took some time kicking melodies around in my head. I knew the basic style I wanted (up-tempo gospel swing) and the chord structure (based on the standard "When the Saints Go Marching In") but melody has always been a challenge for me.
It turns out that I composed the melody one Saturday morning in about 30 minutes, including lyrics. It was completely different from everything that had been going through my head earlier. I even changed up and went to the blues for the verses, which worked great with the "Saints" chorus. I just convinced myself to get it done that morning. By the end of the afternoon I had the choir arrangement finished. It turned out very nicely though unfortunately the recording equipment at the actual event was subpar so the live take sounds worse than something recorded on a Gramophone. But it was great live and in person.
So yes, I absolutely believe that this kind of thing can motivate musicians to create great things. I might have even participated in it myself to get a demo CD I've been wanting to do out the door but I can't afford a good preamp at the moment.
Slashdot Mirror
Well, let's see:
You:
Me quoting you:
Seems a pretty accurate sense of what you said.
Again, you:
And me quoting you:
Again, your original sentence concentrated on the perceived commercial value of the project, not the fact that it's a research project. You intended to shift the debate from one of exploring the design space in a relatively unfettered manner to whether the project serves some narrow economic interest. You did the same when comparing TRIPS to existing commercial architectures when running media applications, ignoring the actual future problems that TRIPS is designed to overcome. I simply focused on the argument you actually made.
Are you now unwilling to stand behind that argument? There's nothing wrong with changing your argument but there is something wrong with doing so while trying to obfuscate the change in debate with attacks.
From your other posts I would guess that you are (loosely) associated with the project / people on it.You would guess wrong.
Please back up your claim about breaking code up into packets (preferably with a paper citation), because if they have done that I would like to read it.Perhaps you would care to examine the TRIPS publications page where there are several papers on the compiler available.
As someone who did his PhD in an area that has no (foreseeable) commercial application I don't really need an explanation of this from you.Yale Patt, is that you?!?
Nod to you for referencing transport-triggered architectures and TRIPS' relationship to them. Good job!
But I must disagree with you about scheduling. Scheduling is HARD. It's hard because there are lots of unknowns like memory latency and how the OOO engine will actually execute the code. The TRIPS compiler doesn't so much schedule the operations themselves as it does the communication. It has to worry about things like routing distance and so on. Yes, a naive implementation would be straightforward but doing something intelligent is often much more difficult.
Thanks for the reference to your work. I'll definitely take a look.
No, it isn't. The TRIPS group has done some really interesting things with compilers, for example. They've managed to have the compiler break up code into packets and schedule them on the processor array so that dependencies flow nicely across the grid. That is not an easy problem to tackle. This is very good research.
That's not the point of research. The point of research is to explore problems no one has tackled before, of course always with an eye toward future technology trends.
Which brings up a good question: what are they counting as ILP? It's been a while since I read the papers so I'll have to go and double-check, but if they're counting predicated instructions that get squashed in their "instructions simultaneously executing" number, it's not a fair comparison to current processors which report IPC as actually instructions that do work.
That said, IIRC TRIPS uses some interesting cache technology to help with speculation and relies on the compiler much more than a standard OOO architecture. If you're running the right kind of codes, there's a rather large amount of parallelism out there to exploit. Even if you're NOT running the right kinds of codes, you can still get a surprising amount of parallelism by removing the stack pointer if your branch prediction (or predication) and latency tolerance is good enough. I don't recall whether TRIPS makes use of this available parallelism or not.
Doug was one of the students who worked on the MultiScalar project at Wisconsin over a decade ago. Lots of good research out of that project and TRIPS bears more than a passing resemblance to MultiScalar. But, of course, it's different. The compilation model is much simpler and in the interim between the projects there have been a number of advances in compiler/architecture cooperation.
TRIPS is definitely not a vector processor. As another poster said, it's close to a dataflow machine, but it also has some of the characteristics of fine-grained multithreading. I'd place it in the spectrum as an architecture engineered to be as close to dataflow as possible while still being practical.
Back when I read the papers, when they had simulator results, it struck me that the individual scalar processors were pretty aggressive, very out-of-order, etc. Complex and difficult to design. I am very curious about what they were able to do in actual silicon. Looks like it's time to read some more papers. :)
Uh huh. That's a good way to put out a "prediction" and be able to claim plausible deniability when it doesn't happen. Time to put up.
Come on now, you're blaming an entire belief system for this tragedy?
This guy did nothing recognizable as Christian, regardless of what he said. Yes, there are despicable people who call themselves "Christian." In fact scripture itself tells us to expect them. Although, let's be VERY clear, it is not a "future telling" set of documents!
Judeo-Christian prophecy is quite different and apart from mere "fortune telling." It is at the heart of the faith tradition, a challenging of society to be something better than it is. There are many, many, many, MANY Christians and people of other faiths doing good work, prophetic work, for real justice in the world. Ghandi was a prophet. Dr. King was a prophet.
I can't speak for other religions (though I suspect their followers would agree with me), but Christian faith is not at all guilt-ridden and does not promote self-loathing. Yes, many people teach it that way and that's tragic. But a good faith community will lift one up, not put one down. I've been in good communities and bad ones. I tend to stay with the good ones.
The core teaching of Christianity is quite simple: God cares about human beings, who are inherently good. It is our responsibility to use our gifts and talents to make the world a better place for everyone.
Now what's threatening in that?
Real ID was passed back as part of an "essential" Iraq funding package. As such it had no separate committee hearings. And at the time, guess who controlled Congress?
Many states are opposed to this not due to privacy concerns but simply because it's another unfunded federal mandate. Minnesota estimates it will cost the state $31 million over five years. Total national costs have ballooned to something like $17 billion. Congress allocated all of $40 million to pay for it.
I know it would take some work for me to produce a certified birth certificate or passport to get a license.
It will be interesting to see what happens when boarding restrictions, etc. go into effect. If this law isn't killed outright at the federal level, I believe it will be effectively ignored in many situations.
More info and an opinion piece:
http://www.startribune.com/587/story/1110277.html http://www.startribune.com/561/story/1119732.htmlIt's not that much of a pig. As a compiler target, it's actually quite nice, except for the occasional register constraints (shift in *CX, etc.). Squeezing performance out can be a pain because of TIMTOWTDI. That makes instruction selection, register allocation and scheduling interact in weird and fascinating ways.
The ModR/M encoding is responsible for the limited register set, but at the time the ISA was created, it was a reasonable tradeoff. X86-64 fixed some of those problems at the expense of more prefixes.
The main problem with the ISA is hardware decoding. Icaches are a pain because you've got to find instruction boundaries. I can imagine that prefixes make the combinational logic rather interesting. Probably a lot of power goes into the decode part of the chip.
But from a programmer-level view, the ISA is actually fairly nice.
The only reason suburbs appear to be less expensive places to live than cities is because we hide the cost of living there. We build the freeways, sewers and other utilities needed for suburban homes but don't reflect that cost in home prices or property taxes. We subsidize the oil going into our cars. The Federal Housing Administration gave grants for suburban construction post-WWII but not for redeveloping cities. Mass transit support has been decimated, making it even harder for those in the cities to get to jobs.
When you learn the history of Detroit, you find its story repeated all over the country. Very prosperous minority neighborhoods were plowed over by the interstate system, destroying an entire economy in the process. We lost untold cultural treasures simply because our racism made it easy to destroy the homes and businesses of a large portion of our community.
Once the tax base is removed from the city, services decrease, poverty concentrates and crime tends to increase. Why does crime increase? Poor schools give young people no hope. The lack of tax base means that urban schools are at a severe disadvantage to suburban schools. The lack of jobs means that young people find other ways to occupy their time.
Fear, ignorance and prejudice causes concentrated poverty because forces manipulate those fears to divide us. Realtors steer whites to white communities, banks redline parts of town, making it more difficult to obtain loans in certain areas.
But even so, people are choosing to live in cities again. Sprawl has gone so far overboard that even with all of the cost hiding, people are spending more money on transportation from the exurbs than they do on their mortgages. I believe we're going to see a radical shift in our living patterns over the next 50 years and we had better begin investing in our infrastructure (e.g. mass transit) to make the transition as smooth as possible.
You've got it exactly backward. Cities declined because people moved out. All sorts of policies were put into place that drove white flight: Federal Housing Administration policy that rewarded new suburban construction over rehabilitation, redlining by lenders, steering by realtors, etc. Some of this stuff still goes on today. Suburbia arose due to deliberate government policy. Robert Moses was the instigator. It wasn't an accident.
Watch this fantastic documentary (parts I & II) about Detroit to learn more. It is not at all the case that cities necessarily decline into wastelands. In many parts of Europe, for example, the concentrated poverty is in suburbia (remember the Paris suburb riots?), which is of course its own problem. Concentrated poverty is the real issue, not cities vs. suburbs. It's our fear, ignorance and prejudice that causes it.
Yes. I had a conversation with an Intel engineer back in 1999 or 2000 stating exactly this.
madhatter is right. Intel made power consumption a key part of their strategy and it is absolutely the correct one. Even in the server market, power is a huge deal. The power to run a datacenter is becoming an increasingly large part of the overall cost of providing services. Heck, even in the HPC market, we're seeing customers demand flops/watt.
AMD had better address this soon or they're going to get eaten alive. Right now, Hypertransport is what's keeping them in the game. When CSI comes along, AMD had better have an answer.
That's a common misperception. It's a lie put forth by the anti-transit crowd to stop investment. The truth is that Atlanta is putting big money into transit, as is Dallas, Denver, Los Angeles, Phoenix and a whole lot of other places that one might consider highly sprawled.
The fact is that practically everywhere there has been major investment in transit, there has been major increase in transit usage. People use transit when it is convenient. That means there needs to be enough service for people to get to work, to the doctor, to the grocery and so on and be certain they can get back without waiting half an hour for a transfer.
Here in the Twin Cities, transit ridership is at a 22-year high, despite 6 years of fare increases and service cuts over the last 7 years. Our bus and rail system is pathetic, but usage is still rising. There is huge demand for transit here and ordinary people (i.e. not lobbyists) are practically knocking down doors at the State Capitol to get funding increased.
Intercity rail is picking up steam in the U.S. For example, the Midwest Regional Rail Initiative would build a network of high-speed rail connecting major cities in the midwest. Amtrak ridership is rising, along with Amtrak revenue.
Transit is a big part of how we will handle the energy and environmental crises we face. Yes the Bush administration proposes to cut $350 million annually from the Federal Transit Administration (FTA). The FTA itself is well-known for how it uses its rulemaking to kill rail projects across the country. I was recently in D.C. talking to our Minnesota delegation about this. Chairman Oberstar will hold oversight hearings and look into this.
We're seeing the start of a major shift in our transportation patterns here in the U.S. and it's not just in the big cities. Medium-sized cities and even rural areas are demanding major investment in transit. It's a good thing, too, because the single-occupancy vehicle has run out of gas. The car will and should always be with us but we can no longer sustain the ever-expanding sprawl and freways to support it without investing in alternative modes.
I used to think this way too. Back in the day.
Language is important. It's how we express ourselves. If one is not willing to take care in writing and speaking, he or she may not give the best impression -- and impressions matter.
There's a difference between error-free writing and proper editing to fix mistakes. No one writes error-free. Editors exist to catch these types of problems.
But to have the attitude going into the process that spelling and grammar don't matter is to isolate oneself in a small group of people that is taken less seriously by the rest of the world.
You've all got it wrong. REAL ID isn't about terrorism, it's about immigration.
The law exists for one reason: some people want to see other people denied basic human rights and opportunity. While the law has scary slippery-slope scenarios, the real danger is what it will do the minute it goes into action: brand some people as lesser than others.
Amen! It's amazing to me the number of people who don't understand this. TIMTOWTDI is the worst aspect of Perl. Well, maybe not the very worst. I'd give that to the OO extensions (and I'm a long-time C++ fan).
TIMTOWTDI pretty much guarantees that no one else will be able to understand your code. It's great for being "clever" and for stroking your ego by impressing your friends with your one-liners, but that won't fly in software that has to be maintained.
Perl is fantastic for what it was originally intended to do: process text. Perl's use of regular expressions is right on. Perl was never meant to be used as a database, for large applications or anything beyond quick & dirty tools.
I use Perl a lot to process reports and create tables for spreadsheets and graphing programs. It works great for that kind of thing. But I would never use it in a large and complex project.
It's fairly easy to tell the differences between two performances of the same piece. Is the pedaling the same? How about dynamics? Tempo? How is rubato used? Are the accents played similarly?
I'm by no means an expert on classical music (traditional jazz is more my thing) but even I can tell the difference between someone who is technically proficient but lacks interpretation and a real pro who understands how the whole piece fits together.
My guess is that the copied performances were not all that well known. Even if a performer is well-known, it's not necessarily the case that all of his or her performances are of such quality that they are universally recognized.
Thank you! I switched to Lilypond several months ago and never looked back. It is so much more flexible than Finale and its ilk due to the fact that it isn't constrained by a graphical representation. I also find that writing music in text is a lot faster than point-and-click or even recording and going back to adjust all of the quantizing problems.
I love the ability to use music variables to hold repeating sequences. I love the programmability (even better with the new streams model). It's extremely easy to write parts for each instrument and mix-and-match them into different scores. I find that, for example, some people in choirs like to see the full SATB parts in a traditional two-staff layout, others prefer a four-staff layout while some prefer just to see their own part. The pianist really wants to see the SATB put on a standard two-staff piano score. No problem with lilypond, I can tailor the presentation to each individual choir member if I wish.
And it makes beautiful engravings, too!
In my opinion, Lilypond completely outclasses commercial and proprietary music scoring software.
Some off-the-wall future consumer things to consider:
I think things get particularly interesting thinking about many-core in handheld devices. Software radio could be tremendously useful there. Route planning using a network of GPS-enabled handheld devices would be cool. Background searching could be used for a lot of things.
100 cores is not massively parallel. The kind of scaling we're talking about is much higher. Think thousands of cores each with hundreds of threads.
This is the kind of scaling that weather centers are just starting to reach today. It's the kind of scaling that will require a radical rethinking of how consumer software is designed and what tools we need to make that design process easier.
In this world, software is king. You won't care who your chip vendor is. You'll care who provides your compiler, debugger, performance analysis tools and other such things.
No, it isn't. It's classic Grover Norquist anti-tax, anti-government rhetoric packaged up in populist clothing. We levy taxes for a very important reason: they fund the services necessary for us to maintain our society and prosper.
It's not insightful to blather on about government inefficiency. You've got to show where it can be improved. More often than not, government programs are running on barebones budgets. There's no fat to cut.
I'm sick of right-wingers fanning the flames of fear and isolation by claiming that there's not enough to go around and that we each have to fight to get our own. I'm sick of those who say we're too poor to meet the needs of everyone in this country. We're the richest country in the history of the world, for God's sake!
Taxes on internet and mail-order purchases are long overdue. I always try to buy locally so that my city, county and state can get just a little more of the funding they need to operate. But I know that not everyone is thinking about that. I've worked up at the Minnesota state legislature (on my own dime and my own time), I've talked with legislators and I've had discussions with staff. I know about the needs of the people in our state and how they are not being met because there are certain elements that want to perpetuate the myth of scarcity for their own benefit.
It's incredibly selfish to whine about a small tax increase when there are homeless children going without food, when low-income people can't get to a job because there's no bus and when funding for neighborhood libraries is being slashed. Doubly so when you complain about having to pay that extra $20 for your shiny new flat-panel display.
The sad thing is, most people buy into the rhetoric without ever understanding or trying to understand what the real situation is. People are aware that something is not right with the way our country is headed. When I point out that the nonsense anti-government and, ultimately, anti-community fear-laced rhetoric is at the root of the problem, they begin to understand and support raising new taxes for services they want.
We make choices. So far, we've chosen to leave people behind in a big way. It's an immoral choice. But it's the choice we've made. We can eradicate poverty and provide opportunity for all. We just have to choose to do it. That means choosing to raise revenue.
Governor Pawlenty, I make a good salary. Tax me a little more so that those less fortunate can have a chance at the same life that I enjoy. Please.
She's not talking about Word. She's talking about ISV codes. Like NASTRAN. Her point is that these codes are not written to scale to thousands of cores. And she's right.
GPGPU and video games are not the same as bioinformatics or particle physics. These are heavy-duty codes, some of which have a very hard time scaling to large numbers of processors. Custom codes are usually much better about this than ISV codes.
Her point is that to use these new architectures, the software stack has to get much better. She's spot on. Certainly PGI and PathScale have not shown yet that they can keep up. I believe they will get there eventually, but it may be much later than we'd like. We will see.
Right on. Deadlines really help too. Last fall we were planning a public meeting for the community organizing group I work with. I told the planning team I was going to compose a choir piece for the event. I had a month to do it and had never written anything for a full choir before.
I took some time kicking melodies around in my head. I knew the basic style I wanted (up-tempo gospel swing) and the chord structure (based on the standard "When the Saints Go Marching In") but melody has always been a challenge for me.
It turns out that I composed the melody one Saturday morning in about 30 minutes, including lyrics. It was completely different from everything that had been going through my head earlier. I even changed up and went to the blues for the verses, which worked great with the "Saints" chorus. I just convinced myself to get it done that morning. By the end of the afternoon I had the choir arrangement finished. It turned out very nicely though unfortunately the recording equipment at the actual event was subpar so the live take sounds worse than something recorded on a Gramophone. But it was great live and in person.
So yes, I absolutely believe that this kind of thing can motivate musicians to create great things. I might have even participated in it myself to get a demo CD I've been wanting to do out the door but I can't afford a good preamp at the moment.