Nvidia's "Tier 0" Program Is Behind The Origin PC Debacle.

[LifesCompanion]

He's absolutely right though, he's saying it like it is.

What happens when something scandelous leaks ? attack the credibility of the leaker... Edward Snowden anyone...

I'm not going to discuss your point regarding Snowden (I don't know enough on the matter) - and WTH happened to my b ) thing? 

 

I digress, the issue I have with all/most of his articles are that he seems to be utterly biased and lacks any form of sources/credibility. If you're going to make a point and want people to believe it, there needs to be evidence, for example, the PCWorldarticle gave some quotes and evidence/statistics, the SemiAccurate article gave none what so ever. 

 

Until there is an official statement that clarifies this, I'm going to give nVidia and Origin the benefit of doubt (innocent until proven otherwise) - to be perfectly honest, just like Omid said:

 

 

Origin had their own gripe with AMD, this was just the final push for them. NV seems to have just gotten burned in the crossfire that would have inevitably come out between Origin and AMD in the future.

 

(I see what you did there Omid, sneaky...) 

[winny3141]

Disregard this post, wrong section

[LifesCompanion]

Found an article (courtesy of HotHardware) that sort of elaborates on the idea that Origin had some beef with AMD and that the recent events were the final draw: http://hothardware.com/News/Why-Everyone-Should-Chill-Out-About-Origin-PC-Dropping-AMD-GPUs/

[LAwLz]

I was excited when I saw this because I thought we finally had proof of all these conspiracy theories but nope, still just Charlie spewing the same old anti-Nvidia propaganda without posting any proof or source whatsoever. Seriously, stop using SemiAccurate as a source. Most mature forums has already. I don't want to wait like 5 years for the people on this site to realize how unreliable and shifty SemiAccurate is.

My secret inside sources said that AMD is paying reviewers to say good things about the 200 series and they actually perform really badly and has lots of issues. My sources are from within AMD so I can't tell you who they are, you're just going to have to take my word for it.

See how ridiculously stupid that sounds? That's what Charlie sounds like to anyone not blinded by their love for AMD. Look critically at things, please. Don't just trust accusations without evidence because you wish for the accusations to be true.

[gastew15]

I wonder if AMD could potential sue for slander if those were nvidia's true intentions. 

Even if they won I'm sure they would get screwed over like they did with Intel.

[Immort]

I was excited when I saw this because I thought we finally had proof of all these conspiracy theories but nope, still just Charlie spewing the same old anti-Nvidia propaganda without posting any proof or source whatsoever. Seriously, stop using SemiAccurate as a source. Most mature forums has already. I don't want to wait like 5 years for the people on this site to realize how unreliable and shifty SemiAccurate is.

My secret inside sources said that AMD is paying reviewers to say good things about the 200 series and they actually perform really badly and has lots of issues. My sources are from within AMD so I can't tell you who they are, you're just going to have to take my word for it.

See how ridiculously stupid that sounds? That's what Charlie sounds like to anyone not blinded by their love for AMD. Look critically at things, please. Don't just trust accusations without evidence because you wish for the accusations to be true.

Those are serious allegations towards Charlie, if you bother to read his work from time to time you'd realize that he's just as harsh on any other company as he is right now on Nvidia, what would you expect if you were a journalist and found out that Nvidia is doing all of this crap, I think you might be the one who's a bit too blinded by your love for Nvidia to look at things critically.

[afa7336]

Those are serious allegations towards Charlie, if you bother to read his work from time to time you'd realize that he's just as harsh on any other company as he is right now on Nvidia, what would you expect if you were a journalist and found out that Nvidia is doing all of this crap, I think you might be the one who's a bit too blinded by your love for Nvidia to look at things critically.

 

Is it your natural reaction to accuse anyone that doesn't agree of being biased? It seems like that is a pattern.

 

What do you make all these other sites that don't believe these rumors by semiaccurate?

[unidentical]

That article was horribly written. A clinic in sensationalist journalism.

 

Until there is a court decision, there's nothing 'unquestionable' about this (quite the opposite, in fact).

 

If nVidia did engage in these sort of dealings, then they should be held accountable, I suspect some people will be fired but you can bet that if this was done, the Board bloody well knew it, but there won't have been an email chain.

 

I suspect Origin will suffer the most in this, even if the allegations prove false, they've lost face.

[LAwLz]

Those are serious allegations towards Charlie, if you bother to read his work from time to time you'd realize that he's just as harsh on any other company as he is right now on Nvidia, what would you expect if you were a journalist and found out that Nvidia is doing all of this crap, I think you might be the one who's a bit too blinded by your love for Nvidia to look at things critically.

Charlie has a history of having a big grudge against Nvidia, and posting nonsense articles (1.7% yield, woodscrews etc) about them. Just because he is harsh against other companies as well does not mean he isn't anti-Nvidia.

I would not expect anyone to trust me without any source, that's for sure. Ever heard of "Extraordinary claims require extraordinary evidence" before? Or "innocent until proven guilty"? If Charlie's accusations are correct, then this is a matter of breaking the law (anti competitive behavior) and trust me, "I got a secret source" won't hold up in court, and neither should it in serious arguments either.

The problem with these types of accusations is that they are impossible to disprove. I can't "prove" that Charlie is wrong, anymore than I can "prove" that Santa isn't real. It is always up to the person making the claim to prove that he/she is correct, not the opposing side to prove that he/she is wrong. Charlie has so far not posted a single evidence that proves that he is correct.

 

And no, I am far from a blind fanboy. I am simply looking critically at things like these (which is why I require solid evidence and not just "well I got evidence but I can't show them to anyone, you just have to take my word for it") in an unbiased way. I would be saying exactly the same thing if the opposite had happened, if AMD had "paid" Origin to get rid of Nvidia. I got a feeling you would be on my side if the situation was the reverse though.

[Mooshi]

Welp, looks like the next card will be AMD.

[MetroP]

Charlie has a history of having a big grudge against Nvidia, and posting nonsense articles (1.7% yield, woodscrews etc) about them. Just because he is harsh against other companies as well does not mean he isn't anti-Nvidia.

I would not expect anyone to trust me without any source, that's for sure. Ever heard of "Extraordinary claims require extraordinary evidence" before? Or "innocent until proven guilty"? If Charlie's accusations are correct, then this is a matter of breaking the law (anti competitive behavior) and trust me, "I got a secret source" won't hold up in court, and neither should it in serious arguments either.

These aren't extraordinary claims, not at all.

Nvidia have been known to have very little ethics in their business practices, much like Dell and Microsoft.

Charlie never bad mouths any company, what he says is always true, he's been telling the inside story of what's going on forever now, I wouldn't stop believeing him for two reasons, #1 he's always been accurate #2 he gains nothing from his allegations against Nvidia #3 he exposed many less than ethical practices before so I wouldn't stop taking his word now, just because it happens Nvidia is the one in the crap hole.

 

[afa7336]

These aren't extraordinary claims, not at all.

Nvidia have been known to have very little ethics in their business practices, much like Dell and Microsoft.

Charlie never bad mouths any company, what he says is always true, he's been telling the inside story of what's going on forever now, I wouldn't stop believeing him for two reasons, #1 he's always been accurate #2 he gains nothing from his allegations against Nvidia #3 he exposed many less than ethical practices before so I wouldn't stop taking his word now, just because it happens Nvidia is the one in the crap hole.

 

 

So, since he is always right, I got to point out two wonderful headlines just for you.

 

"Get a Nvidia GTX770 free! Bios flash and GTX680 not included but required." and "Why would Nvidia release an unlocked GTX680 to GTX770 BIOS?"

 

Since your claim is that he is always correct, explain this lovely piece of sh...stuff, by this author.

[Kuzma]

The problem is Maxwell isn't anywhere near ready yet lol :3  But yeah, on paper it looks to be a very powerful lineup, i'm liking the idea of the combined memory already since DDR4 will bring much higher frequencies to system RAM, offsetting some of the performance gaps between DDR3 and GDDR5

but latency is rising in scale with the frequency rise so

[Gabriel]

So we have Nvidia fighting with everything they have to secure whatever market share they have left, throwing spaghetti at the wall to see what sticks & then Linus says Nvidia is holding back, do you realize how ridiculous that sounds ?

Nvidia could fuck up the whole market though just put titan for 50$ and its GG

[LAwLz]

These aren't extraordinary claims, not at all.

Nvidia have been known to have very little ethics in their business practices, much like Dell and Microsoft.

Charlie never bad mouths any company, what he says is always true, he's been telling the inside story of what's going on forever now, I wouldn't stop believeing him for two reasons, #1 he's always been accurate #2 he gains nothing from his allegations against Nvidia #3 he exposed many less than ethical practices before so I wouldn't stop taking his word now, just because it happens Nvidia is the one in the crap hole.

Holy crap how out of touch with this world are you? Like someone said before, Charlie is like a broken clock. He screams bad things about Nvidia 12 times a day and is correct 2 of them.

 

Charlie usually has 2 inaccurate posts for each 1 accurate post, but people forget the inaccurate ones or he uses logical fallacies to make them impossible to disprove (like in this case when he just goes "I got a source but I won't tell anybody who it is, so you just have to take my word for it").

Want some inaccurate predictions from Charlie? How about when he said Apple would switch to ARM processors in their laptops by 2013? Did that happen? No...

How about when he said Apple would switch to AMD GPUs? Well let's see... Early 2009 models had Nvidia. The iMac then changed to AMD GPUs and then back to Nvidia again but the Macbook Pros continued to use Nvidia and Intel graphics (until 2-3 years later when they briefly changed to AMD graphics in the Macbook Pro which caused quite a bit of issues when it tried to switch back and forth between dedicated and integrated GPUs, and then switched back to Nvidia again).

Remember when Charlie said Nvidia was going to abandon the mid-range and high end market after the GeForce 200 series because they "can't compete with ATI"? Yeah, after that they released the 480... So obviously he was very wrong on that prediction.

Oh and how can we forget when Charlie said the GT300 yields were under 2%? Wanna know what his source was? A Japanese blog post he ran through Google translate, which translated the article wrong (he thought it said 9 engineering samples worked which is where he got the 1.7% yield from, but in fact it said the engineering samples was going to be completed in September). ephud also explains why Charlie is wrong in this thread.

 

So I've already proven that Charlie is not always correct (in several major instances where he has been wrong above). I got several more but I don't want to give any more pageviews to his inaccurate website which is mostly hit phishing with sensational headlines and next to no hard evidence.

[LAwLz]

but latency is rising in scale with the frequency rise so

Ehh... Where did you get that from? As far as I know, that's completely untrue.

[Kuzma]

Ehh... Where did you get that from? As far as I know, that's completely untrue.

Look at the transistion between DDR2 and DDR3 , the latencies rose with the frenquenies - albeit slightly offset.

[LAwLz]

Look at the transistion between DDR2 and DDR3 , the latencies rose with the frenquenies - albeit slightly offset.

I don't think you understand how CL is measured. A 1600MHz CL9 kit has higher latency than a 2133MHz kit with CL9. That's because the CL is measured in how many clock cycles it takes, so the shorter the clock cycle (which becomes shorter with a higher Hz) the less latency each #CL adds.

So a 1333MHz CL8 kit will have higher latency than a 2133MHz kit with CL9 (I think. Didn't actually calculate how many ms latency the both kits will have but you get the point).

[Kuzma]

I don't think you understand how CL is measured. A 1600MHz CL9 kit has higher latency than a 2133MHz kit with CL9. That's because the CL is measured in how many clock cycles it takes, so the shorter the clock cycle (which becomes shorter with a higher Hz) the less latency each #CL adds.

So a 1333MHz CL8 kit will have higher latency than a 2133MHz kit with CL9 (I think. Didn't actually calculate how many ms latency the both kits will have but you get the point).

However a CL8 800Mhz kit with have lower latency than a CL19 1066Mhz kit which is what we were looking at when DDR3 initially became a thing so expect similar things to happen this time.

[LAwLz]

However a CL8 800Mhz kit with have lower latency than a CL19 1066Mhz kit which is what we were looking at when DDR3 initially became a thing so expect similar things to happen this time.

1066MHz CL19? Are you sure that even existed? Because when I was looking for memory for the LGA 1366 (before it was released) we had a ton of 1600MHz CL9 kits. I don't even think 1066MHz CL19 is in the DDR3 spec. Got any source?

 

As for the DDR4 latency. The test sample from Samsung seems to be CL13. Let's assume that it just barely makes the DDR4 specs with a frequency of 2133MHz. That means it would have a latency of ~12.2 nanoseconds.

Your standard 1600MHz CL9 kit has a latency of 11.25 nanoseconds. So the engineering samples of DDR4 isn't that bad as far as latency goes. It's a ~8% increase in latency but a ~33% increase in bandwidth, on an unfinished product. I expect the latency on DDR4 to be about the same as we got on DDR3, if not lower, like a year after it is first released.

[exyia]

the internet has made "journalism" such a complete joke

 

idiots like this are free to post garbage articles - and all that matters is how many page views they get for ads. terrible writing, no credible sources, no documentation, no concise/focused argument or debate - oh doesn't matter, advertisers are happy. thanks for the page view

[Kuzma]

1066MHz CL19? Are you sure that even existed? Because when I was looking for memory for the LGA 1366 (before it was released) we had a ton of 1600MHz CL9 kits. I don't even think 1066MHz CL19 is in the DDR3 spec. Got any source?

I mean CL9 sorry

[LAwLz]

I mean CL9 sorry

Ahh that sounds better. The 800MHz kit at CL8 would still have higher latency than 1033Mhz at CL9 though.

 

CL8 at 800MHz = 20ns latency.

CL9 at 1066MHz = 16.9ns latency.

 

 

To calculate latency you do this (we will use 1600MHz CL9 as an example):

First you need the real clock rate of the RAM. RAM is double data rate so manufacturer usually advertises the "effective clock rate" which is twice as high as the real clock rate (400 MHz and can send two signals for each clock = 800MHz of normal performance, so they call it 800MHz). So take the advertised frequency and cut it in half. 1600 / 2 = 800. We have to remember that we just calculated what it would be in Hz and we should use MHz, so take 800 and multiply it with 1000000 (to get from Hz to MHz).

 

Now we need to calculate how long 1 cycle takes. Take 1 and divide it by your clock speed, in this case 800000000 (which is our real life clock rate of 1600MHz RAM, in Hz). 1 / 800000000 = 0.00000000125. Each cycle takes 0.00000000125 seconds to complete. Covert seconds to nanoseconds and you get 1.25.

So each cycle on 1600MHz RAM takes 1.25 nanoseconds.

 

Simply multiply that with the CL (in this case 9) and you get the total latency (for the first memory column, the other columns will have higher latency). in this case it is 1.25 * 9 = 11.25 nanoseconds latency.

[Kuzma]

Ahh that sounds better. The 800MHz kit at CL8 would still have higher latency than 1033Mhz at CL9 though.

 

CL8 at 800MHz = 20ns latency.

CL9 at 1066MHz = 16.9ns latency.

 

 

To calculate latency you do this (we will use 1600MHz CL9 as an example):

First you need the real clock rate of the RAM. RAM is double data rate so manufacturer usually advertises the "effective clock rate" which is twice as high as the real clock rate (400 MHz and can send two signals for each clock = 800MHz of normal performance, so they call it 800MHz). So take the advertised frequency and cut it in half. 1600 / 2 = 800. We have to remember that we just calculated what it would be in Hz and we should use MHz, so take 800 and multiply it with 1000000 (to get from Hz to MHz).

 

Now we need to calculate how long 1 cycle takes. Take 1 and divide it by your clock speed, in this case 800000000 (which is our real life clock rate of 1600MHz RAM, in Hz). 1 / 800000000 = 0.00000000125. Each cycle takes 0.00000000125 seconds to complete. Covert seconds to nanoseconds and you get 1.25.

So each cycle on 1600MHz RAM takes 1.25 nanoseconds.

 

Simply multiply that with the CL (in this case 9) and you get the total latency (for the first memory column, the other columns will have higher latency). in this case it is 1.25 * 9 = 11.25 nanoseconds latency.

i was wrong okay well roll on the DDR4!

[LAwLz]

i was wrong okay well roll on the DDR4!

Yeah most people seem to be confused about latency vs frequency when it comes to RAM and it's easy to assume that a higher number always means higher latency. But now you know, and knowing is half the battle.