Re:IBM PCs compared extremely poorly with Amigas
on
The Amiga Turns 25
·
· Score: 5, Insightful
Complete and utter rubbish.
The effect that you are refering to only happened in the business market, and the home market was where Commodore made most of their money and sales.
What killed the Amiga was stagnation. Sure it was way ahead of the competition when it was released, but it didn't improve enough, quickly enough. By the mid-90s the Amiga was competing against chunky 256-colour display and faster processors.
Doom killed the Amiga. Comanche killed the Amiga. Every step that the PC took towards being a commodity marketplace for hardware killed the Amiga.
And by the time the Voodoo was released it was already dead.
Not only does this make sense, but sometimes later never comes.
I rip DVDs straight to a drive array as plain isos. The original plan was to get around to re-encoding the video as h264 when the drives were full / I could be bothered. I don't buy new drives very often but the drive array is still growing faster than it is filling.
The next hike is 2GB drives to replace some of the oldest in the array now that they are down to 100 quid. As this is prompted by errors starting to show up on an old drive rather than a need for space it seems possible that cheap storage has exceeded my needs for storage space.
However, again, it really doesn't matter at the end of the day, because the fact is PC game sales aren't doing well, and if you want to pretend otherwise fine, but it wont change that fact.
I didn't really make this clear at the beginning, although I've mentioned in in a reply to another poster on this thread: I don't disagree with you that PC sales are poor. My only disagreement was that you pulled some figures from your arse and presented them as fact. You're not the only one that does it. Far from it. But I find it strange that so many people accept the impending death of PC games as a solid fact without any solid evidence.
Just to be really clear: I'm not arguing that PC gaming is alive and healthy - just that no evidence (other than anecdote and opinion) has been presented to prove it. One thing that I do find interesting though is that the impended doom of the PC gaming industry was forecast at this point in the console life-cycle last time and it didn't come to pass. Over the life of that cycle I would expect industry focus to shift between consoles and PCs relative to which provides the best eye-candy as that point.
As far the NPD figures go, apologies as I thought I had put a link in the post but I clearly cocked that up. I can't find the page in my history and it was a serious pain in the ass to find the first time. Edit: middle of this page between the pie charts. Also, I should have pointed out the first time that this is the units sold for the top ten games on each platform, so it is not the total but I would not expect the total to be much higher given that the distribution of sales is almost winner-takes-all.
Some interesting tidbits and annecdotes in there, thanks for that. While most of what you say does paint the same picture that Xest had in mind there is still a lack of definite figures. The issue that I had with Xest wasn't really his opinion that PC gaming is dead, it was his statements of fact that were just hot air, and then his second worse attempt to cover it. I'd be quite happy to admit that consoles are killing PC games in term of sales.... but for once I would like to see some definite evidence rather than inflated opinion (on his part) and interesting annecdote (on yours).
That 170k sales figure for MW2 in November is taken from NPD. Yet it was NPDs figures that I quoted as placing the non-digital download part of the PC games market as being on par with each of the major console platforms.
Obviously something is wonky as the same numbers don't simultaneously prove that the PC market is both on par with a major console and a fraction of the size. Do you thinking that picking one months sales for one title might introduce a teeny weeny bit of skew into the data?
Are you really so hard of thinking? I criticized you for pulling figures out of your arse. Your response is even worse. Let me explain it to you in small pieces to increase your chance of understanding.
Units of what exactly? I'm a little puzzled, you're suggesting you can't find annual sales figures, and then suggesting you have some annual sales figures
I don't have access to actual sales figures. But I provided estimates from a reliable source (NPD is recognised as an industry standard for estimates of units sold in the games market and they have access to a lot of figures that the public do not).
So then you provide some figures instead of pulling them out of your own arse. Let's see where they come from shall we:
* Polling end users to find out what games they are currently purchasing and playing * Polling retail partners to find out what games and hardware they are selling * Using statistical trend fitting and historical data for similar games * Studying resell prices to determine consumer demand and inventory level
That taken from http://www.vgchartz.com/methodology.php. Or in other words they pull them out of their ass. So which do I trust more: a market research company trusted by the industry who estimates sales at 20 million based on access to non-public info, or somebody with a fucking Z in their name who ran a phone poll and estimated 100 million units......
Hmmmmm. Keep trying, you may actually find a source to back up your imaginary figures you look hard enough.
Wrong, wrong, wrong wrong. In fact you are so wrong that I'm undoing mods to reply to you.
You take the figures from the summary and then produce some figures from your arse, claim they are bigger and therefore the article is wrong? Who says that Super Mario Kart wii sold 21.3 units in the US in 2009? Your claims are at odds with wikipedia which claims that 22 million copies have been sold world-wide in the two years since launch.
Although I can't find annual sales figures for consoles in 2009 I have at least looked a bit harder than you to find some real figures: NPD sales figures for the US in 2009 show 22.6 million units sold for the Wii, 20.4 million units for the xbox360 and 8.7 million for the PS3.
So the PC market for digital downloads is the same size as the most popular platforms, and the total PC games market is twice the size. Quite the opposite of your conclusions, but then I used real numbers instead of those stored up my arse.
We currently pay about 300kr a month for a 30Mb connection. I think that's about 30euro / 25pounds / $40. We don't get throttled and there are no limits as far as I know. BT tends to max out at 3MB/s on popular torrents, lower than that if the swarm isn't big enough to saturate the line.
There are cheaper packages available, and our ISP goes up to 100Mb/s symmetric.
I've got a 15" MBP from a couple of years ago with a matte screen. My wife has a brand new 13" MBP with glossy. The differences are fairly standard - hers looks a lot more colourful, but it is like looking into a mirror. I find the reflection very distracting and would definitely go for matte when upgrading (if I get the option).
For indoors use I think most people would prefer the glossy look (because the reflections don't bother them so much). For outdoors use I haven't tried the glossy screen in direct sunlight but I know that the matte one remains readable.
Ah, slashdot, the place where the GP clearly read the description of a "tall window producing a radial whitish highlight" and reasoned correctly that the GGP was talking about being in a relatively dark room compared to the submitter's query about direct sunlight near the equator, but the parent failed to notice this and tried to criticise him.
Hint: there are more light levels in the world than "dark" and "bright" and what the GP thinks of as bright is very, very far down the dark scale.
What kind of crack are you smoking? This story is not about the NYT, and the NYT does not have a paywall (registration is free). Did you think that you could make some kind of point by cobbling together words that you felt were related to the story?
Thank you captain obvious. I understood the point that he was making. Now take a deep breath and bearing that in mind reread my post. Do you now suspect that perhaps I was making the point that universal education has a direct benefit to him that is more apparent than the indirect benefits he was talking about?
It could be more tightly related than that, but it depends on how scalable and reliable it is. Rather than predicting an increase in transistor density, Moore actually predicted an increase in transistor density at a fixed price-point. This is the same as a reducing price-point for a given transistor budget.
So one way to increase power would be to assemble processors out of smaller units. If each unit is an IC with a fixed transistor budget and this provides a reliable and scalable way to assemble those ICs into chips then it could have quite a large impact on the growth of processor performance.
No, it's the difference between "efficiency" and what is claimed as "efficient" to get a paper published. That's a really bad citation for AES on GPUs as there is a line of prior work going back to Cook and Cryptographics. In fact that paper is a classic example of getting something into the literature that has already been done. The authors have submitted it to an unrelated conference and failed to cite the relevant work.
If we look at their best figures then throw away the 15x claimed speedup as it doesn't consider memory transfer costs. The 5x speedup is more realistic. The GPU that they use (8800gtx) has 128 stream processors running at 1.35Ghz. The comparison is a PIV running at 3Ghz. Roughly speaking we can compare the cycles taken on each platform as a measure of the work done. The graphics card stream processors perform 57x more clock cycles.
The central workload in AES for high-performance is completely memory bound. The cycles are just used to stage results from memory and perform XOR instructions. So the stream processors only execute the code 5x quicker with 57x more clocks and a huge memory bandwidth advantage that I can't be bothered to look up.
So no, 10x less output per clock is not "efficient" in my book. But if you publish your paper in a crappy unrelated conference then you will get away with it.
But it is not the same kind of maths. Most GPUs support very fast use of single-precision floats. The asymmetric crypto that you use to establish your SSL connection uses very large integers, and the AES that encrypts the stream operates in a finite field. Neither can executed efficiently on a GPU.
That would not be a double-blind study. If the person who is looking after the plants is aware of which ones are shielded it could affect their levels of care for each of them. Using a separate judge would not fix this problem.
But a proper double-blind trial is trivial. Setup a set of "hoods" for the plants. Make half of them shielded and half not. Don't let the person tending the plants know which are which. Do the judging first and then reveal which plants were shielded and which were not afterwards.
Instead of assuming the OP was talking about the iPhone 4G antenna issue think back a little way further. The Santa Barbara generation of MacBook-pros used faulty 8600GTs from nvidia. This was the process shrink that nvidia totally fucked up and shipped millions of faulty GPUs.
Unlike most pc manufacturers who issued recalls straight away Apple hedged for 12 months before announcing that the thousands of failing machines did have a hardware fault and they would replace any model in the line that showed problems within two years.
I say this a little bitterly as my my early 2008 laptop is now 2.5 years old and has stopped using as many PCIe lanes as it should on the GPU which is an early sign of failure. Now remind me again how spotless Apple are?
f(x) = f(x-1)+1 is quite easy to parallelise. Same for f(x) = f(x-1) + f(x-2) although doing so has spawned many research papers.
Would anyone care to name a specific problem that requires computing the nth element of a recursively defined sequence which is hard to do parallelise?
Slashdot Mirror
Complete and utter rubbish.
The effect that you are refering to only happened in the business market, and the home market was where Commodore made most of their money and sales.
What killed the Amiga was stagnation. Sure it was way ahead of the competition when it was released, but it didn't improve enough, quickly enough. By the mid-90s the Amiga was competing against chunky 256-colour display and faster processors.
Doom killed the Amiga. Comanche killed the Amiga. Every step that the PC took towards being a commodity marketplace for hardware killed the Amiga.
And by the time the Voodoo was released it was already dead.
Not only does this make sense, but sometimes later never comes.
I rip DVDs straight to a drive array as plain isos. The original plan was to get around to re-encoding the video as h264 when the drives were full / I could be bothered. I don't buy new drives very often but the drive array is still growing faster than it is filling.
The next hike is 2GB drives to replace some of the oldest in the array now that they are down to 100 quid. As this is prompted by errors starting to show up on an old drive rather than a need for space it seems possible that cheap storage has exceeded my needs for storage space.
That's a bold claim without anything to substantiate it.
Which compiler researchers believe that a competent human will outperform a compiler?
I didn't really make this clear at the beginning, although I've mentioned in in a reply to another poster on this thread: I don't disagree with you that PC sales are poor. My only disagreement was that you pulled some figures from your arse and presented them as fact. You're not the only one that does it. Far from it. But I find it strange that so many people accept the impending death of PC games as a solid fact without any solid evidence.
Just to be really clear: I'm not arguing that PC gaming is alive and healthy - just that no evidence (other than anecdote and opinion) has been presented to prove it. One thing that I do find interesting though is that the impended doom of the PC gaming industry was forecast at this point in the console life-cycle last time and it didn't come to pass. Over the life of that cycle I would expect industry focus to shift between consoles and PCs relative to which provides the best eye-candy as that point.
As far the NPD figures go, apologies as I thought I had put a link in the post but I clearly cocked that up. I can't find the page in my history and it was a serious pain in the ass to find the first time. Edit: middle of this page between the pie charts. Also, I should have pointed out the first time that this is the units sold for the top ten games on each platform, so it is not the total but I would not expect the total to be much higher given that the distribution of sales is almost winner-takes-all.
Some interesting tidbits and annecdotes in there, thanks for that. While most of what you say does paint the same picture that Xest had in mind there is still a lack of definite figures. The issue that I had with Xest wasn't really his opinion that PC gaming is dead, it was his statements of fact that were just hot air, and then his second worse attempt to cover it. I'd be quite happy to admit that consoles are killing PC games in term of sales.... but for once I would like to see some definite evidence rather than inflated opinion (on his part) and interesting annecdote (on yours).
That 170k sales figure for MW2 in November is taken from NPD. Yet it was NPDs figures that I quoted as placing the non-digital download part of the PC games market as being on par with each of the major console platforms.
Obviously something is wonky as the same numbers don't simultaneously prove that the PC market is both on par with a major console and a fraction of the size. Do you thinking that picking one months sales for one title might introduce a teeny weeny bit of skew into the data?
Are you really so hard of thinking? I criticized you for pulling figures out of your arse. Your response is even worse. Let me explain it to you in small pieces to increase your chance of understanding.
I don't have access to actual sales figures. But I provided estimates from a reliable source (NPD is recognised as an industry standard for estimates of units sold in the games market and they have access to a lot of figures that the public do not).
So then you provide some figures instead of pulling them out of your own arse. Let's see where they come from shall we:
That taken from http://www.vgchartz.com/methodology.php. Or in other words they pull them out of their ass. So which do I trust more: a market research company trusted by the industry who estimates sales at 20 million based on access to non-public info, or somebody with a fucking Z in their name who ran a phone poll and estimated 100 million units......
Hmmmmm. Keep trying, you may actually find a source to back up your imaginary figures you look hard enough.
Wrong, wrong, wrong wrong. In fact you are so wrong that I'm undoing mods to reply to you.
You take the figures from the summary and then produce some figures from your arse, claim they are bigger and therefore the article is wrong? Who says that Super Mario Kart wii sold 21.3 units in the US in 2009? Your claims are at odds with wikipedia which claims that 22 million copies have been sold world-wide in the two years since launch.
Although I can't find annual sales figures for consoles in 2009 I have at least looked a bit harder than you to find some real figures: NPD sales figures for the US in 2009 show 22.6 million units sold for the Wii, 20.4 million units for the xbox360 and 8.7 million for the PS3.
So the PC market for digital downloads is the same size as the most popular platforms, and the total PC games market is twice the size. Quite the opposite of your conclusions, but then I used real numbers instead of those stored up my arse.
We currently pay about 300kr a month for a 30Mb connection. I think that's about 30euro / 25pounds / $40. We don't get throttled and there are no limits as far as I know. BT tends to max out at 3MB/s on popular torrents, lower than that if the swarm isn't big enough to saturate the line.
There are cheaper packages available, and our ISP goes up to 100Mb/s symmetric.
Ah, ok. Well in that case let me say WHOOOOSH for you, and then .... I'll get my coat.
I've got a 15" MBP from a couple of years ago with a matte screen. My wife has a brand new 13" MBP with glossy. The differences are fairly standard - hers looks a lot more colourful, but it is like looking into a mirror. I find the reflection very distracting and would definitely go for matte when upgrading (if I get the option).
For indoors use I think most people would prefer the glossy look (because the reflections don't bother them so much). For outdoors use I haven't tried the glossy screen in direct sunlight but I know that the matte one remains readable.
Ah, slashdot, the place where the GP clearly read the description of a "tall window producing a radial whitish highlight" and reasoned correctly that the GGP was talking about being in a relatively dark room compared to the submitter's query about direct sunlight near the equator, but the parent failed to notice this and tried to criticise him.
Hint: there are more light levels in the world than "dark" and "bright" and what the GP thinks of as bright is very, very far down the dark scale.
Err.... WTF?
What kind of crack are you smoking? This story is not about the NYT, and the NYT does not have a paywall (registration is free). Did you think that you could make some kind of point by cobbling together words that you felt were related to the story?
Thank you captain obvious. I understood the point that he was making. Now take a deep breath and bearing that in mind reread my post. Do you now suspect that perhaps I was making the point that universal education has a direct benefit to him that is more apparent than the indirect benefits he was talking about?
Because you don't want to live in a society where your neighbours are uneducated?
It could be more tightly related than that, but it depends on how scalable and reliable it is. Rather than predicting an increase in transistor density, Moore actually predicted an increase in transistor density at a fixed price-point. This is the same as a reducing price-point for a given transistor budget.
So one way to increase power would be to assemble processors out of smaller units. If each unit is an IC with a fixed transistor budget and this provides a reliable and scalable way to assemble those ICs into chips then it could have quite a large impact on the growth of processor performance.
Nah. Just means that they fitted the bumpers before selling it.
No, it's the difference between "efficiency" and what is claimed as "efficient" to get a paper published. That's a really bad citation for AES on GPUs as there is a line of prior work going back to Cook and Cryptographics. In fact that paper is a classic example of getting something into the literature that has already been done. The authors have submitted it to an unrelated conference and failed to cite the relevant work.
If we look at their best figures then throw away the 15x claimed speedup as it doesn't consider memory transfer costs. The 5x speedup is more realistic. The GPU that they use (8800gtx) has 128 stream processors running at 1.35Ghz. The comparison is a PIV running at 3Ghz. Roughly speaking we can compare the cycles taken on each platform as a measure of the work done. The graphics card stream processors perform 57x more clock cycles.
The central workload in AES for high-performance is completely memory bound. The cycles are just used to stage results from memory and perform XOR instructions. So the stream processors only execute the code 5x quicker with 57x more clocks and a huge memory bandwidth advantage that I can't be bothered to look up.
So no, 10x less output per clock is not "efficient" in my book. But if you publish your paper in a crappy unrelated conference then you will get away with it.
But it is not the same kind of maths. Most GPUs support very fast use of single-precision floats. The asymmetric crypto that you use to establish your SSL connection uses very large integers, and the AES that encrypts the stream operates in a finite field. Neither can executed efficiently on a GPU.
Damn. I have modpoints but I can't mod a reply to me funny in this universe.
Where do you go to shoot zombies in 3D in actual reality?
So we can look forward to unlimited* launch?
* Fair usage cap of 25ft
You are quite right that you are not qualified. But on the subject tof US patents being enforced abroad you are also quite wrong.
That would not be a double-blind study. If the person who is looking after the plants is aware of which ones are shielded it could affect their levels of care for each of them. Using a separate judge would not fix this problem.
But a proper double-blind trial is trivial. Setup a set of "hoods" for the plants. Make half of them shielded and half not. Don't let the person tending the plants know which are which. Do the judging first and then reveal which plants were shielded and which were not afterwards.
Come on people, this is really basic shit here.
What a short memory the AC has.
Instead of assuming the OP was talking about the iPhone 4G antenna issue think back a little way further. The Santa Barbara generation of MacBook-pros used faulty 8600GTs from nvidia. This was the process shrink that nvidia totally fucked up and shipped millions of faulty GPUs.
Unlike most pc manufacturers who issued recalls straight away Apple hedged for 12 months before announcing that the thousands of failing machines did have a hardware fault and they would replace any model in the line that showed problems within two years.
I say this a little bitterly as my my early 2008 laptop is now 2.5 years old and has stopped using as many PCIe lanes as it should on the GPU which is an early sign of failure. Now remind me again how spotless Apple are?
Not quite, but a good effort.
f(x) = f(x-1)+1 is quite easy to parallelise.
Same for f(x) = f(x-1) + f(x-2) although doing so has spawned many research papers.
Would anyone care to name a specific problem that requires computing the nth element of a recursively defined sequence which is hard to do parallelise?