Intel Z3770 vs Snapdragon 800 in Kraken 1.1

[Homeles]

If you're going to change your argument to "relative numbers" then yes, they do.

It would behoove you to understand the concepts of "relevance" and "context."

You can't make the simple leap between those two?

Do you really think most tablet buyers are going in to shops and running SPECint on them before making a purchase decision?

Guess that means Intel's graphics get a free pass, since no one will be benching their products to be able to tell the difference.

Why don't you do some work for yourself, google it and come back and tell us.

That would be your job. You're trying to present the case that GPUs are important, but if you can't prove that intensive gaming even makes up a sizable portion of application usage, your argument is null and void.

You really do lack basic logic don't you homeles.

In your most desperate, wettest dreams, perhaps.

What I'm saying is that Apple made the choice to go with big graphics performance in their tablets. If gaming performance wasn't important then why waste a huge amount of die area on graphics? If you can't figure out what that means then that's your problem.

It is important. I have pointed this out multiple times now.

Your failure to pick up on this makes your insult above rather ironic.

Tell me, when are you going to prove that it is the most important thing? When are you going to prove that Intel is royally screwed because their Achilles Heel is sub-par GPU performance?

 

[liahos1]

lets tone it down, this thread too good to get shut down.

Silicon, why you think Tegra 4 got nowhere. Under your logic it has a comparable GPU to s800 but a better cpu? Thoughts?

Also given:

1) INTEL's density advantage in the 2013-2016 time frame
2) Their now jihadist focus on winning in mobile
3) the fact they have the most to lose if they dont win in mobile
4) the low amount of mobile share they have

Is it really so hard to suggest intel has some chance of success here? I'm not talking about 90% of the market overnight, but certainly 10% next year

 

[Homeles]

1) INTEL's density advantage in the 2013-2016 time frame
2) Their now jihadist focus on winning in mobile
3) the fact they have the most to lose if they dont win in mobile
4) the low amount of mobile share they have

Is it really so hard to suggest intel has some chance of success here? I'm not talking about 90% of the market overnight, but certainly 10% next year

I should point out that Intel's transistor performance is head and shoulders above everyone else. Even their 32nm bulk process outperformed IBM/GloFo's 32nm PDSOI.

I don't feel like they have the most to lose -- they're not really a big mobile player, so losing mobile wouldn't really hurt them that badly. They could theoretically cut out mobile SoC development and do just fine (although I am not sure what the profitability of Atom is). They could cut their losses on the fabs that they've built in anticipation for higher demand, and run a very efficient company, still raking in billions every year. They have invested a lot — that's for sure — but if they were to flat out turn around and quit the mobile market right now, they'd still be in the green.

However, it's the goal of businesses to expand. The game isn't so much about what they have to lose; it's about what they have to gain.

Regardless of whether or not Silvermont is a home run or not, it's easy to see that the Silvermont architecture and 22nm will improve Atom's standing against its ARM competitors. I.e., regardless of whether or not Atom is a slam dunk, Intel will gain market share.

 

[SiliconWars]

Guess that means Intel's graphics get a free pass, since no one will be benching their products to be able to tell the difference.

Yeah I mean it's not like you can tell the difference in graphics with your eyes or anything like that. Oh...wait.

That would be your job. You're trying to present the case that GPUs are important, but if you can't prove that intensive gaming even makes up a sizable portion of application usage, your argument is null and void.

The only proof needed is that the industry leader has put a very large amount of die area into graphics. All the real tablet players are going the same way with more graphics performance.

On the flipside we have Intel with small graphics and tiny market share. There does appear to be something of a correlation?

Tell me, when are you going to prove that it is the most important thing?

Where did I say it was?

When are you going to prove that Intel is royally screwed because their Achilles Heel is sub-par GPU performance?

From what I can see, you are the one telling us all that Intel is going to totally dominate tablets by the end of this year.

Let me help you stop making stupid statements homeles.

http://www.digitimes.com/news/a20130812PD217.html

Read, and weep, at the total lack of Bay Trail in any tablet that matters.

 

[Khato]

Another subject that is all too frequently forgotten when it comes to comparisons of graphics between the various smartphone/tablet SoCs is just how different they look compared to what we're used to in the notebook/desktop realm. Which actually has been fine since the rendering methods that have been the norm for PC gaming only scale down so far effectively... but the same is also true, the graphics pipelines used in mobile to date only scale up so far before their glass jaws become unavoidable. Basically, if it were capable of playing a modern PC game, Adreno 330 might well perform worse than Baytrail's anemic 4 EU graphics just because it was never designed to handle that level of complexity. Whereas typical mobile gaming actually plays to Gen7's primary weakness of inadequate texture sampling resources.

On the bright side, hopefully it's only around 2 weeks left before all this speculation will be put to rest with actual reviews.

 

[SiliconWars]

lets tone it down, this thread too good to get shut down.

Silicon, why you think Tegra 4 got nowhere. Under your logic it has a comparable GPU to s800 but a better cpu? Thoughts?

Tegra 4 appears to have power and heat issues. Read this, especially the parts on power and random hangs etc (temps, throttling and power consumption near the bottom of the article) - http://www.notebookcheck.net/Review-Toshiba-Excite-Pro-AT10LE-A-108-Tablet.98828.0.html

Also given:

1) INTEL's density advantage in the 2013-2016 time frame
2) Their now jihadist focus on winning in mobile
3) the fact they have the most to lose if they dont win in mobile
4) the low amount of mobile share they have

Is it really so hard to suggest intel has some chance of success here? I'm not talking about 90% of the market overnight, but certainly 10% next year

It's hard to pinpoint what kind of percentages they'll be looking at winning next year. You have to remember that it *is* a long term thing. This Atom is what they are going to fight mobile with for years.

That said, Qualcomm basically say that they expect to overtake Intel in tablets - http://hexus.net/business/news/comp...-snapdragon-800-will-build-tablet-lead-intel/

This appears to have already happened, and then some. Some people seem to think that Qualcomm and the rest are just standing around like bowling pins waiting on Intel. They aren't. Not by a long shot. I would not be surprised if Qualcomm increases both phone and tablet market share over Intel next year.

Anyway, we're not supposed to be discussing the market in this thread, so let's keep to tech discussion as much as we can.

 

[Homeles]

Yeah I mean it's not like you can tell the difference in graphics with your eyes or anything like that. Oh...wait.

You can tell the performance difference between two different tablets running 2D games?

The only proof needed is that the industry leader has put a very large amount of die area into graphics. All the real tablet players are going the same way with more graphics performance.

Intel didn't design Silvermont with the hopes that Apple would adopt it, buddy.

Intel's graphics will still be at a pretty respectable level. Why do they need to be the top player, exactly?

On the flipside we have Intel with small graphics and tiny market share. There does appear to be something of a correlation?

On the other side of things, we see that AMD has big graphics and an even smaller market share. There doesn't appear to be something of a correlation?

Where did I say it was?

You wouldn't be so desperately invested in this topic otherwise.

From what I can see, you are the one telling us all that Intel is going to totally dominate tablets by the end of this year.

No, that's not what I'm saying either. They're going to have the best SoC. I can't predict market adoption.

Let me help you stop making stupid statements homeles.

Aww, does this mean you're going to be my little helper? How cute!

http://www.digitimes.com/news/a20130812PD217.html

Read, and weep, at the total lack of Bay Trail in any tablet that matters.

That's a very complete list you have there. Regardless of your pathetic piece of evidence, where have I stated that I expect Intel to get some grand-slam OEM win?

I'm more interested in smartphones anyway.

This Atom is what they are going to fight mobile with for years.

Keep telling yourself that. You'll be awfully surprised when Airmont lands next year.

 

[SlimFan]

I think the question would revolve around whether or not there are games out there that are too heavy to play on an Adreno 320? Wouldn't that be the list of games that are too heavy for Bay Trail since the estimates are that they are similar performance?

For Android I would imagine there are no games that fall into that category. For Windows 8, there are many, many games that fall into that category, none of which can run on Windows RT anyway.

 

[liahos1]

Tegra 4 appears to have power and heat issues. Read this, especially the parts on power and random hangs etc - http://www.notebookcheck.net/Review-Toshiba-Excite-Pro-AT10LE-A-108-Tablet.98828.0.html

It's hard to pinpoint what kind of percentages they'll be looking at winning next year. You have to remember that it *is* a long term thing. This Atom is what they are going to fight mobile with for years.

That said, Qualcomm basically say that they expect to overtake Intel in tablets - http://hexus.net/business/news/comp...-snapdragon-800-will-build-tablet-lead-intel/

This appears to have already happened, and then some. Some people seem to think that Qualcomm and the rest are just standing around like bowling pins waiting on Intel. They aren't. Not by a long shot. I would not be surprised if Qualcomm increases both phone and tablet market share over Intel next year.

Anyway, we're not supposed to be discussing the market in this thread, so let's keep to tech discussion as much as we can.

fair point. how could qualcomm possibly increase its share in mobile against intel next year? Intel had basically no phone share and all the tablet share they had was in windows 8 tablets.

 

[SiliconWars]

You can tell the performance difference between two different tablets running 2D games?

Where are your numbers showing that 2d gaming is what most people are doing on their tablets?

Intel didn't design Silvermont with the hopes that Apple would adopt it, buddy.

Which begs the question, who DO they hope adopts it?

Intel's graphics will still be at a pretty respectable level. Why do they need to be the top player, exactly?

Intel's best graphics will be respectable. All A15's and Kraits, and hell even Tegra 4 will have "respectable" CPU performance.

On the other side of things, we see that AMD has big graphics and an even smaller market share. There doesn't appear to be something of a correlation?

We could then extrapolate that the reason for this is that AMD doesn't give a monkey's about the tablet market. And we'd be right in saying that. Why would AMD care about selling $20 chips for the tablet market when the Kabini die they are based on sells for 3x that in laptops?

You wouldn't be so desperately invested in this topic otherwise.

Oh so now you're reduced to making guesses as to what my motives are? Do yourself a favour homeles, quit while you're still only a mile behind.

No, that's not what I'm saying either. They're going to have the best SoC. I can't predict market adoption.

How can it be the best SoC when it has 70% slower graphics and no modem?

There, homeles. There is your best SoC, by any metric that matters that there is the best SoC on the market for the next year.

Aww, does this mean you're going to be my little helper? How cute!
That's a very complete list you have there. Regardless of your pathetic piece of evidence, where have I stated that I expect Intel to get some grand-slam OEM win?

So what are you expecting? That intel builds their own like Nvidia has been forced to do? Good luck with that lol.

I'm more interested in smartphones anyway.

Current Intel market share, 0.2% or so? I must admit, I *do* believe that will improve over the coming year.

Keep telling yourself that. You'll be awfully surprised when Airmont lands next year.

Be sure to wake me up when that happens.

 

[Homeles]

Where are your numbers showing that 2d gaming is what most people are doing on their tablets?

Really, you do need to do the world a favor and learn the concept of "burden of proof."

Which begs the question, who DO they hope adopts it?

Ah, a misusage of the phrase "begging the question." Allow me to educate you: http://begthequestion.info/

In direct response to your asinine question: any one of the dozens of players in the tablet industry and smartphone industry.

Intel's best graphics will be respectable. All A15's and Kraits, and hell even Tegra 4 will have "respectable" CPU performance.

Sure. But unlike the A15s, Intel-powered tablets won't run the risk of turning into molten slag and will have far superior battery life.

And, er, your knowledge of the English language isn't aiding you.

When one says "All _____, even _____ will be better/on/par/etc.," the noun(s) following "even _____ " is supposed to be lesser than the noun(s) following "All _____."

Example with correct usage (with added tongue-in-cheek humor):
"All Intel processors, even Pentium 4s, are better than AMD's processors."

We could then extrapolate that the reason for this is that AMD doesn't give a monkey's about the tablet market. And we'd be right in saying that.

Good grief you are clueless. So tell me why Dirk Meyer was fired again?

Why would AMD care about selling $20 chips for the tablet market when the Kabini die they are based on sells for 3x that in laptops?

http://www.idc.com/getdoc.jsp?containerId=prUS24129713

Yep. AMD doesn't care. Good one.

Oh so now you're reduced to making guesses as to what my motives are? Do yourself a favour homeles, quit while you're still only a mile behind.

http://en.wikipedia.org/wiki/Dunning–Kruger_effect

Should I start tallying how many times you've been wrong vs. how many times I've been wrong?

How can it be the best SoC when it has 70% slower graphics and no modem?

Apparently the concept of "overall best" is one that eludes you.

As far as the modem goes, the lack of integration doesn't put them at as large of a disadvantage as you're making it out to be.

http://anandtech.com/show/7234/intel-talks-about-multimode-lte-modems-xmm7160-and-beyond

So what are you expecting? That intel builds their own like Nvidia has been forced to do? Good luck with that lol.

Why do you operate in a world where everything is black and white? Why does Intel have to capture 100% of the tablet market next year, or else Silvermont is a failure? That is essentially what you are making this out to be.

This isn't life or death. If Intel doesn't jump from their roughly 6.75% market share to 40%, it's not the end of the world.

Current Intel market share, 0.3% or so? I must admit, I *do* believe that will improve over the coming year.

So do I. Also, your implication that I care about Saltwell is rather embarrassing.

Meanwhile in the tablet world: ~6.75%.

http://www.eetimes.com/author.asp?section_id=36&doc_id=1287258

~6.75% and growing rapidly, that is:
http://www.businessinsider.com/intel-says-tablet-market-share-is-growing-2013-6

Be sure to wake me up when that happens.

Most of this forum would prefer if you stayed asleep.

 

[sushiwarrior]

In direct response to your asinine question: any one of the dozens of players in the tablet industry and smartphone industry.
Sure. But unlike the A15s, Intel-powered tablets won't run the risk of turning into molten slag and will have far superior battery life.

And, er, your knowledge of the English language isn't aiding you.

When one says "All _____, even _____ will be better/on/par/etc.," the noun(s) following "even _____ " is supposed to be lesser than the noun(s) following "All _____."

Example with correct usage (with added tongue-in-cheek humor):
"All Intel processors, even Pentium 4s, are better than AMD's processors."
Good grief you are clueless. So tell me why Dirk Meyer was fired again?
http://www.idc.com/getdoc.jsp?containerId=prUS24129713

Yep. AMD doesn't care. Good one.
http://en.wikipedia.org/wiki/Dunning–Kruger_effect

Should I start tallying how many times you've been wrong vs. how many times I've been wrong?
Apparently the concept of "overall best" is one that eludes you.

As far as the modem goes, the lack of integration doesn't put them at as large of a disadvantage as you're making it out to be.

http://anandtech.com/show/7234/intel-talks-about-multimode-lte-modems-xmm7160-and-beyond
Why do you operate in a world where everything is black and white? Why does Intel have to capture 100% of the tablet market next year, or else Silvermont is a failure? That is essentially what you are making this out to be.

This isn't life or death. If Intel doesn't jump from their roughly 6.75% market share to 40%, it's not the end of the world.
So do I. Also, your implication that I care about Saltwell is rather embarrassing.

Meanwhile in the tablet world: ~6.75%.

http://www.eetimes.com/author.asp?section_id=36&doc_id=1287258

~6.75% and growing rapidly, that is:
http://www.businessinsider.com/intel-says-tablet-market-share-is-growing-2013-6
Most of this forum would prefer if you stayed asleep.

1. I don't see A15's as turning anything into molten slag, and don't talk preemptively about Atom's power/heat.
2. Tegra 4 delivers inferior "day to day" CPU performance (perhaps 4+1 design?), even though it performs well in benchmarks - see the review.
3. No modem essentially means no real smartphone SoC.
4. He said smartphone SoC, because you said smartphone, now you're changing your mind and saying tablet marketshare instead - please stop changing the terms to inflate your assertions.
5. Intel is betting the farm on mobile, so to speak, so yes, if they don't gain significant market share it is a failed bet (doubtful to cause major harm to the company overall, but certainly not GOOD for them).
6. Many members appreciate Silicon, keep your opinion to yourself :whiste:

 

[TuxDave]

6. Many members appreciate Silicon, keep your opinion to yourself :whiste:

What if I appreciate Homeles' opinion? Then what?

Oh the bickering in this thread. At the end of the day, marketing from both sides will confuse the consumer on what the best choice is and people will end up buying a product based on the OEM's brand name. /just my uninformed opinion

 

[SiliconWars]

Really, you do need to do the world a favor and learn the concept of "burden of proof."

Yes, seeing as you are the one banging on about 2d...? Yup, that's right Homeles.

Ah, a misusage of the phrase "begging the question." Allow me to educate you: http://begthequestion.info/

You really are desperate. I could waste time explaining this to you but seeing as it's a tech forum I'm not going to bother.

In direct response to your asinine question: any one of the dozens of players in the tablet industry and smartphone industry.

Perhaps you can extrapolate by telling us how the huge amount of cash Intel has pumped into mobile, and continues to lose quarter over quarter, is going to be recouped? I'd love to hear your in-depth analysis of how Intel is going to manage this. Really I would. I'm all ears, not literally so don't bother looking for another "misuse of phrase".

Sure. But unlike the A15s, Intel-powered tablets won't run the risk of turning into molten slag and will have far superior battery life.

This is based on? See homeles, for somebody who constantly insists on others proving their points, you sure do talk a lot and never seem to show anything.

And, er, your knowledge of the English language isn't aiding you.

When one says "All _____, even _____ will be better/on/par/etc.," the noun(s) following "even _____ " is supposed to be lesser than the noun(s) following "All _____."

Example with correct usage (with added tongue-in-cheek humor):
"All Intel processors, even Pentium 4s, are better than AMD's processors."

Just sad, please give it up.

Good grief you are clueless. So tell me why Dirk Meyer was fired again?

http://www.idc.com/getdoc.jsp?containerId=prUS24129713

Yep. AMD doesn't care. Good one.

Not being privy to AMD's board decisions, I really don't know. Sure I heard the rumours. Why do I get the feeling that if I looked hard enough I'd find evidence of you saying it was because of Bulldozer or something else though? The thing about Meyer's dismissal is that Intel fanboys generally use whatever reason fits at the time as proof. Like with most things regarding AMD, frankly.

http://en.wikipedia.org/wiki/Dunning%E2%80%93Kruger_effect

Really not interested in your reading habits homeles. Nobody is. This is a tech forum not a psychology forum.

Apparently the concept of "overall best" is one that eludes you.

No it's quite apparent. Telling us that Intel has the best SoC is a bit like going to a Nobel prize dinner night and telling everybody that Usain Bolt is the best human being alive. You're getting a similar reaction here btw, you just can't see everybody laughing at you.

As far as the modem goes, the lack of integration doesn't put them at as large of a disadvantage as you're making it out to be.

http://anandtech.com/show/7234/intel-talks-about-multimode-lte-modems-xmm7160-and-beyond

No homeles, as I continue to say, it's the lack of design wins and market share that is the disadvantage.

Why do you operate in a world where everything is black and white? Why does Intel have to capture 100% of the tablet market next year, or else Silvermont is a failure? That is essentially what you are making this out to be.

Untrue. I simply don't listen to nonsense about Intel KO'ing everybody in this space. It's a joke, honestly it's just so lowbrow to believe it. The argument is always the same "intel has the most amazing process blah blah". Yep, they've always had it, and never managed a knockout blow yet.

This isn't life or death. If Intel doesn't jump from their roughly 6.75% market share to 40%, it's not the end of the world.

I've never believed that the end of Intel would be the end of the world. It's true Intel doesn't NEED this - but they need to make sure they don't lose what they already have. That's the point homeles. That's the point.

~6.75%. Off by over an order of magnitude. I'd say nice try, but let's be honest — you couldn't have made a worse guess.

http://www.eetimes.com/author.asp?section_id=36&doc_id=1287258

Tablets != phones.

Most of this forum would prefer if you stayed asleep.

Almost certainly. Then again seeing you being taken apart like this surely makes it worth it for many of them.

 

[Nothingness]

This isn't life or death. If Intel doesn't jump from their roughly 6.75% market share to 40%, it's not the end of the world.
So do I. Also, your implication that I care about Saltwell is rather embarrassing.

Meanwhile in the tablet world: ~6.75%.

http://www.eetimes.com/author.asp?section_id=36&doc_id=1287258

How do you get 6.75% from the article you link? From the 7.4% of Windows minus RT?

~6.75% and growing rapidly, that is:
http://www.businessinsider.com/intel-says-tablet-market-share-is-growing-2013-6

You first link uses a study from Strategy Analytics. Their numbers for Q2 are much worse for Windows: https://www.strategyanalytics.com/default.aspx?a0=5403&mod=pressreleaseviewer
Yes, Windows went down from 7.4% to 4.3% (even the number of units went down). Given that there was no Intel Android tablet in Q2, Intel would at best have 4.3% market share.

Anyway, as you wrote, it can't get worse for Intel, they can only gain market shares :biggrin:

All of that said, I do think there are some nuances in Javascript (as I've always said) that make it not really THAT representative for benches. And while I hate to complain about pretty much every benchmark used I'm going to have to give the usual aside even about SPECInt. It's a good benchmark but it's indisputable that ICC generates much better code for it on x86 than GCC (VIA did a paper that showed exactly that, ~20% performance advantage on Nano) and that Intel is going to use ICC for any promotional numbers.

For people interested, the study is here: http://www.via.com.tw/en/downloads/whitepapers/processors/NanoX2_whitepaper_201107.pdf
icc gains 24% over gcc on SPEC INT 2k. That's much more than the gain I got on any of my program with icc (oh well except for that particular part of nbench :whiste.

That doesn't mean the comparison is illegitimate or they're doing anything wrong, but I still contend that very little Android software will use ICC except for those that are sponsored by Intel.

Intel has just released icc for Android, and guess what? It's currently free.
http://software.intel.com/en-us/c-compiler-android

I'll play with that compiler against Android gcc on SPEC 2k to see what speedup it brings.

 

[liahos1]

s600 didnt have an integrate modem this year. didnt prevent qcom from doing quite well. Apple doesnt use an integrated modem. Samsung exynos products dont use integrated modems. When tegra 2 came out qcom had an integrated 3g soc, tegra 2 did quite well.

Intel supplied the HSPA variant baseband to the S4 which most ppl forget. The die area of S800 > 33% greater than s600 due to inclusion of lte modem. i'd like to see the overall tdp of the s800 w/ and w/o integration. I'd also like to see a like for like TDP with a Merrifield and separate xmm7160/7260.

carriers want baseband competition for QCOM. This is from the CTO at verizon. i think its too early to write intel off!

i'd also like to see how many tablets are going to ship with LTE vs consumers buying these at best buy as compute companions.

This could be an endless debate. But the debate has shifted. Awhile back many ppl (at least ppl i talked to) said that arm was somehow magically/structurally superior to x86. proven false. They said that intel could never build a capable Soc. proven false. said that software compatibility is lacking. proven false. they said that windows on arm would destroy intel. prove false. they said that intel would never focus on mobile because of its financial impact. proven false. They said intel would never be capable without lte. has lte now.

Now the argument is GPU is only as good or slightly better than an s600 so they are doomed.
Intel only supports 15 lte bands. Intel is doomed.
Intel doesnt have an integrated baseband. Intel is doomed.

but clearly there is a roadmap for these things to improve. 14nm brings a mountain of die budget qcom wont have. Volte baseband by end of year. integration still an open question.

what happens if pc sales stabilize next year and starts to get share in tablets/smartphones. will intel still be doomed? peoples perceptions can shift so quickly. some ppl only focus on current data points and try to extrapolate into infinity. it'll be an interesting next few years.

 

[mavere]

For people interested, the study is here: http://www.via.com.tw/en/downloads/w...per_201107.pdf
icc gains 24% over gcc on SPEC INT 2k. That's much more than the gain I got on any of my program with icc (oh well except for that particular part of nbench ).

Christ what a silly paper. They specifically DISABLED -O# optimizations for their tests, which makes those benchmarks the equivalent of the Special Olympics: interesting only if you want to compare handicaps. No one compiles their release binaries with disabled optimizations.

I'm not even defending Intel here. I'm defending GCC, as a 20%+ performance disadvantage in something like SPEC is a serious insult.

 

[Nothingness]

Christ what a silly paper. They specifically DISABLED -O# optimizations for their tests, which makes those benchmarks the equivalent of the Special Olympics: interesting only if you want to compare handicaps. No one compiles their release binaries with disabled optimizations.

You're right, no one compiles without optimization except on Android NDK where default flags for ARM disable many optimizations by optimizing for size, while for Intel it's full opt...

In that particular case, -O3 was used by VIA, even PGO was used. See page 15 of the document.

I'm not even defending Intel here. I'm defending GCC, as a 20%+ performance disadvantage in something like SPEC is a serious insult.

icc is known as SPEC compiler :biggrin:

 

[Exophase]

Christ what a silly paper. They specifically DISABLED -O# optimizations for their tests, which makes those benchmarks the equivalent of the Special Olympics: interesting only if you want to compare handicaps. No one compiles their release binaries with disabled optimizations.

I'm not even defending Intel here. I'm defending GCC, as a 20%+ performance disadvantage in something like SPEC is a serious insult.

Don't be absurd. Just because they're not in the peak flags doesn't mean they weren't there. Go cross reference these scores with Intel published scores for Atom and tell me that you really think they were done at -O0.

You can see these flags in page 15. This is the relevant excerpt:

[FONT=sans-serif]##########[/FONT]
[FONT=sans-serif]######################################################[/FONT]
[FONT=sans-serif]# Baseline Tuning Flags[/FONT]
[FONT=sans-serif]################################################################[/FONT]

[FONT=sans-serif]int=base=default=default:[/FONT]

[FONT=sans-serif]COPTIMIZE[/FONT][FONT=sans-serif]= [/FONT][FONT=sans-serif]-[/FONT][FONT=sans-serif]O3 $(BASE_FLAGS)[/FONT]
[FONT=sans-serif]CXXOPTIMIZE[/FONT][FONT=sans-serif]= [/FONT][FONT=sans-serif]-[/FONT][FONT=sans-serif]O3 $(BASE_FLAGS)[/FONT]
[FONT=sans-serif]fp=base=default=defaul[/FONT][FONT=sans-serif]t:[/FONT]

[FONT=sans-serif]FOPTIMIZE[/FONT][FONT=sans-serif]= [/FONT][FONT=sans-serif]-[/FONT][FONT=sans-serif]O3 $(BASE_FLAGS)[/FONT]
[FONT=sans-serif]F77OPTIMIZE[/FONT][FONT=sans-serif]= [/FONT][FONT=sans-serif]-[/FONT][FONT=sans-serif]O3 $(BASE_FLAGS)[/FONT]
[FONT=sans-serif]COPTIMIZE[/FONT][FONT=sans-serif]= [/FONT][FONT=sans-serif]-[/FONT][FONT=sans-serif]O3 $(BASE_FLAGS)[/FONT]

 

[mavere]

Hmm, you're correct. My mistake.

 

[Arachnotronic]

It's representative of SPECint, yes. Notice they didn't use SPECfp?

Notice that after using S800 and Tegra 4 for SPECint, they then used the S4 and Tegra 3 for the other benchmarks (including Sysmark? lol)

I guess that's because you don't like the Geekbench 2 results we already have?

In Geekbench 2, leaked BYT samples outperform S800 in Integer and non-buggy FPU. But then again, I can't blame you for not doing the extensive due diligence with respect to this matter that I and others have, which is a good part of why I find your "Intel's Bay Trail will suck" comments face-palm worthy.

Bay Trail puts Intel in the game, by being best-in-class on CPU, having a competitive (albeit not T4/S800 class) GPU, and the right IP in the SoC to be truly competitive. It won't do what Conroe did to AMD, but it will get Intel a solid amount of share from which to build upon in Cherry Trail and then Willow Trail.

 

[SiliconWars]

In Geekbench 2, leaked BYT samples outperform S800 in Integer and non-buggy FPU. But then again, I can't blame you for not doing the extensive due diligence with respect to this matter that I and others have, which is a good part of why I find your "Intel's Bay Trail will suck" comments face-palm worthy.

I guess we'll see. The truth is almost certainly somewhere in-between points of view, as always. Nowhere did I ever say that Bay Trail will suck btw.

Bay Trail puts Intel in the game, by being best-in-class on CPU, having a competitive (albeit not T4/S800 class) GPU, and the right IP in the SoC to be truly competitive. It won't do what Conroe did to AMD, but it will get Intel a solid amount of share from which to build upon in Cherry Trail and then Willow Trail.

Where is this share coming from? It's not any of the new iPad's (Apple), it's not the Kindle Fire (Qualcomm), it's not the Nexus 7 (Qualcomm) or Galaxy tabs (Samsung and Qualcomm supposedly) either.

All these huge brand tablets are launching at the same time Intel is launching Bay Trail. I really can't see how anything is going to change in Intels favour when they are up against that.

Look (North American website impressions last year but still kinda relevant to the point) - http://chitika.com/insights/2012/november-tablet-market-update

If you're not in the iPad, Nexus, Galaxy or Kindle then what are you in?

 

[bullzz]

@liahos1
interesting. i always thought s600 had the integrated modem. funny that s600 is specific to phones and they are more likely to use LTE than tablets

also funny seeing ppl bashing intel for releasing a chip with a weak GPU but think AMD is totally oblivious to all of this competition

 

[bullzz]

@SiliconWars

i guess we have already discussed this. the chip has not even been officially released yet. give it a month
true, they will not be in any of these products. but will be in most windows tablets. windows OEMs dont have many choices

 

[Homeles]

1. I don't see A15's as turning anything into molten slag, and don't talk preemptively about Atom's power/heat.
2. Tegra 4 delivers inferior "day to day" CPU performance (perhaps 4+1 design?), even though it performs well in benchmarks - see the review.
3. No modem essentially means no real smartphone SoC.
4. He said smartphone SoC, because you said smartphone, now you're changing your mind and saying tablet marketshare instead - please stop changing the terms to inflate your assertions.
5. Intel is betting the farm on mobile, so to speak, so yes, if they don't gain significant market share it is a failed bet (doubtful to cause major harm to the company overall, but certainly not GOOD for them).
6. Many members appreciate Silicon, keep your opinion to yourself :whiste:

1. I'm talking preemptively about Atom's power/heat — sorry if you don't like it. Also, check Tegra 4's heat issues that SiW linked.
2. And Silvermont will be better.
3. Okay.
4. Yes, I misread what he had posted — shame on me. Regardless, it his numbers still have no relevance to me.
5. Exactly, it's not life and death, so why is SiW making it out to be?
6. And the vast majority of those members would be in better off over at AMDZone.

_________________________

Stealing sushi's format.

1. Yes, seeing as you are the one banging on about 2d...? Yup, that's right Homeles.
2.You really are desperate. I could waste time explaining this to you but seeing as it's a tech forum I'm not going to bother.
3.Perhaps you can extrapolate by telling us how the huge amount of cash Intel has pumped into mobile, and continues to lose quarter over quarter, is going to be recouped? I'd love to hear your in-depth analysis of how Intel is going to manage this. Really I would. I'm all ears, not literally so don't bother looking for another "misuse of phrase".
4. This is based on? See homeles, for somebody who constantly insists on others proving their points, you sure do talk a lot and never seem to show anything.
5. Just sad, please give it up.
6. Not being privy to AMD's board decisions, I really don't know. Sure I heard the rumours. Why do I get the feeling that if I looked hard enough I'd find evidence of you saying it was because of Bulldozer or something else though? The thing about Meyer's dismissal is that Intel fanboys generally use whatever reason fits at the time as proof. Like with most things regarding AMD, frankly.
7. Really not interested in your reading habits homeles. Nobody is. This is a tech forum not a psychology forum.
8. No it's quite apparent. Telling us that Intel has the best SoC is a bit like going to a Nobel prize dinner night and telling everybody that Usain Bolt is the best human being alive. You're getting a similar reaction here btw, you just can't see everybody laughing at you.
9. No homeles, as I continue to say, it's the lack of design wins and market share that is the disadvantage.
10. Untrue. I simply don't listen to nonsense about Intel KO'ing everybody in this space. It's a joke, honestly it's just so lowbrow to believe it. The argument is always the same "intel has the most amazing process blah blah". Yep, they've always had it, and never managed a knockout blow yet.
11.I've never believed that the end of Intel would be the end of the world. It's true Intel doesn't NEED this - but they need to make sure they don't lose what they already have. That's the point homeles. That's the point.
12. Tablets != phones.
13. Almost certainly. Then again seeing you being taken apart like this surely makes it worth it for many of them.

1. Until you prove that Intel's Gen 7 graphics aren't enough for most people, you have no argument. Period. In order to do this, you need to prove that 3D graphics are a relevant concern.
2. Translation: "You got me. I quit."
3. Please explain to me why you are worth enough of my time for me to bother with producing such numbers. Here's some proof of Intel's future success.
4. The very link with Tegra 4 you posted. Oops, you shouldn't have shown me that.
5. I sincerely apologize for being better with the English language than you.
6. So uh, if AMD doesn't care about tablets, what's Temash all about? You don't exactly spend millions of R&D on something you don't care about.
7. Oh, okay. You like to glorify ignorance. I see now — you're one of those people.
8. Whatever makes you feel warm and cheery inside!
9. Gee, it's almost as if you think you know something special. It's not like every tech journalist hasn't commented on this already.
10. That might be because their current mobile arch is 5 years old — keep up with the times.
11. All signs point in the opposite direction. Intel will be growing, not shrinking. I don't think anyone's surprised that their market share has tapered off since last year.
12. I misread your post. Try reading people's edits next time.
13. I can go grab that tally for you if you'd like...

______________________

How do you get 6.75% from the article you link? From the 7.4% of Windows minus RT?

0.9 * 0.074 = 0.0666

Mistyped 0.074 as 0.075 for my 6.75% number.

You first link uses a study from Strategy Analytics. Their numbers for Q2 are much worse for Windows: https://www.strategyanalytics.com/default.aspx?a0=5403&mod=pressreleaseviewer
Yes, Windows went down from 7.4% to 4.3% (even the number of units went down). Given that there was no Intel Android tablet in Q2, Intel would at best have 4.3% market share.

Pretty damn good considering that Atom's running on a 5 year old μarch, and it's their first year.