How Did Anand Let This Get Published?

[crashtestdummy]

The sentiments in the article are simultaneously correct and misplaced. Haswell is not really much of an upgrade as far as desktop systems go, and it's true that the reason for that is Intel is feeling very little pressure at the high end.

What's missing is that this is not due to AMD being lacking, but rather to the lack of a profitable market there. Desktops are quickly become a niche market, and most gaming systems are GPU rather than CPU limited anyway. So even if there were a sizable boost in processing power, it would only be of use to a very, very small population. AMD has no interest in pursuing that population, either.

 

[Arachnotronic]

Get over it OP. Haswell sucks. So does Ivy. But SANDY is still sexy sexy SANDY!! (I just bought a 4770k for work but SHHHHHH!!!!)

Yeah, I'd much rather have this aging Sandy Bridge than a new Haswell...not. If I weren't waiting to get an IVB-E, I'd be all over Haswell.

 

[BallaTheFeared]

How about latest Handbrake (0.9.9.5530)? You can re-code this to mp4 using High Profile preset and note the times, to make it comparable to other systems.

NB: HB uses some SSE 4.2/AVX instructions, so "ancient" processors will be somewhat penalized. Keep in mind.

This isn't going to happen, the video download is brutally slow.

 

[Magic Carpet]

Where is the thread with those results

I'll make a thread later

This isn't going to happen, the video download is brutally slow.

Pick a different mirror/torrent? Any x1080 MP4.

 

[grimpr]

lol

 

[Enigmoid]

I've found that AT notebook reviews are somewhat lacking.

What would be nice....

1. Power Consumption and Clocks. (Can the chip run full turbo under load? Anand remarked that turbo seems to be in favor of the cpu in richland. Good. Now tell us the clockspeed of the cpu/igp in trinity and richland). Power consumption varies greatly across notebooks but given that AT has the a10-4600m GX60 and now the a10-5750m GX 60 --same model, power consumption should be given). (Lots of times its reported that the notebook is throttling-- Good but what are the throttling clocks; like in the MSI GT70 Dragon: are we just not getting turbo or are we below turbo? What levels of stress are required to induce throttling?).

2. More tests: Winrar, adobe, compiling tests, etc. Generally AT seems to skimp on the benchmarks for mobile.

3. Broader system comparison. How does the a10-5750m compare to other notebooks that are NOT ultrabooks. Why are the only notebooks in the comparison kabini (which is a completely different segment), or ULV. Compare it to some other 35 watt part. Would really be nice to see a comparison to HD 4000 non-ULV or HD 4600 (throw the desktop HD 4600 in there for comparison as we can subtract off about 10-20% for a mobile chip at 1150 mhz; the data they already have).

4. Better numerical measurements such as noise levels.

 

[nyker96]

I must say I agree with the reviewer that Intel has no high end competition to drive them in making faster desktop parts. They are now mostly focused on mobile optimized parts, as they should, that's where the money is right now.

We can probably expect Intel's upgrade cycle to slow down year over year from this point on. I believe they will continue to refresh the graphics core part of the cpu just to keep pushing the GPU end on the mobiles but CPU core portion development will slow down as there don't seem to be any demand for such powerful CPU in any applications.

What I think is that unless someone can come up w/ a killer app for desktop CPU, there will be little need for a dedicated powerful machine to do our daily chores. The CPU end will continue to slow down.

 

[TidusZ]

Kudos to Anandtech for telling it like it is and not sugarcoating the fact that 2700k=3770k=4770k

 

[Arachnotronic]

Kudos to Anandtech for telling it like it is and not sugarcoating the fact that 2700k=3770k=4770k

No, Anand Shimpi was positive enough on 4770k...

 

[SiliconWars]

No, Anand Shimpi was positive enough on 4770k...

Considering the underwhelming reaction elsewhere, you might want to consider that the problem may be with Anand and not Dustin.

 

[Pandamonia]

From the recent AMD A10 mobile APU review,



I'm sorry, but while I do agree that desktop Haswell is underwhelming for those looking to upgrade from Ivy Bridge, it seems like complete [travesty] that this reviewer would actually say that Intel is "coasting" on Haswell when it was very clearly a major step in the right direction...mobile.

Haswell ULT brought some serious battery life gains, and Haswell in general brought us ~10% IPC increase in legacy code + some seriously cool new instruction set extensions. Broadwell should be good for mobile/AIO devices where it will be in BGA form and probably will focus much more on IGP improvements than CPU improvements anyway, so it'd be pointless to bring it to LGA.

Thoughts?


No profanity in the tech forums, please.

Moderator jvroig

I think your clear bias and total defence of a company is worse than anything written by any reviewer. At least the reviewer has his own opinion and the reasons why.

In my opinion your just one of those guys who will defend Intel like someone who works for them. i think its pretty sad to be honest.

 

[Blitzvogel]

I won't complain as much about the IPC as I would complain that Intel hasn't been pushing for desktop octacores until recently. Sure IPC is nice, but with further gravitation towards multithreading it would be nice for not only system capability but longevity too. Plenty of enthusiasts would've gone for it.

 

[Grooveriding]

Desktop Haswell is a joke. Particularly in the context of how little it brings compared to SB if you account for overclocking.

The only interesting chip to come out of Haswell will be Haswell-E when it brings eight cores to the enthusiast lineup and solder between the die and IHS.

Kudos for the reviewer calling it like it is.

 

[blackened23]

No, Anand Shimpi was positive enough on 4770k...

Intel17, respectfully, I agree with you regarding Haswell and think it is a fine chip. I don't see what people have against it as it maintains the IPC increase just as IVB did - and the mobile version is simply a groundbreaking product. Haswell ULV absolutely delivered everything intel promised in spades.

However with that being said, reviewers can and will be subjective at times - and the review you linked was not even for the Haswell. If you want to see the 4770k blasted, there are quite a few websites that simply were way worse in terms of harsh language against the Haswell. Tomshardware comes to mind, but there were many more. Again, reviewers can be somewhat subjective even if you don't agree. That makes for interesting reading at times, or frustrating if you disagree. I think we both disagree with the reviewer in this case, but that doesn't make his opinion invalid, he's an opinionated individual just like everyone else. If reviewers weren't opinionated at times, they would all come across as a canned press release.

Bottom line, even though I disagree, I see nothing wrong with the language used in the review. I would rather have an opinionated reviewer rather than one that won't ever "commit" to giving a true opinion - that would make for very boring reading.

 

[Arachnotronic]

I think your clear bias and total defence of a company is worse than anything written by any reviewer. At least the reviewer has his own opinion and the reasons why.

In my opinion your just one of those guys who will defend Intel like someone who works for them. i think its pretty sad to be honest.

I don't work for Intel.

 

[Pandamonia]

Considering the underwhelming reaction elsewhere, you might want to consider that the problem may be with Anand and not Dustin.

Anand is an intel fanboi.

Its pretty clear if you read his reviews.

I like reading the articles but Anand is likely more worried about losing access to intel if he came on too strong about intel poor showings.

Lets face it, Atom is a failure, IGP is still a failure, performance in desktop has stalled for 3 years. Intels chipsets have lacked features which should have been included but took years to be added. They have taken features out and pushed them up to higher priced CPU's and chipsets. Most of the time this stuff gets brushed over. You only notice when you realise its suddenly missing.

Generally if you read AT though intel can do no wrong. Im glad someone has called intel out.

 

[Pandamonia]

I don't work for Intel.

If you said that you did then i could understand the way you are.

But since you have no horse in this race then i cant explain your behaviour to a company which doesnt give a monkey about you.

 

[Gikaseixas]

Fact = Haswell on desktop is not that appealing. It's not AMD's fault either, it's just Intel being Intel circa Northwood era

 

[aigomorla]

lulz..

im happy to see im not the only one getting called traitor for saying haswell is failwell.

Honestly tho... if guys like me.. who used to be very deep routed in the intel camp can straight up and say intel fubard, doesnt that speak much?
True i havent been active in a while... yet i would hope i built some brownie points with you guys and intel to know im at least somewhat reputable.

Haswell has been nothing but face palms upon face palms....
And trust me... i have been waiting and waiting for it, only to realize i should of jumped on the 3970K bandwagon last year. :\
Now i need to get IVY-E.... and possibly just wait til broadcomm-E while skipping haswell-e entirely.

 

[blackened23]

Fact = Haswell on desktop is not that appealing. It's not AMD's fault either, it's just Intel being Intel circa Northwood era

It's not appealing ONLY if you insist on upgrading your CPU every 12 months. On it's own merit, the 4770k is a damn fine chip IMHO, it maintains the exact same IPC increase over it's predecessor as IVB did. I see nothing wrong with that.

I can't think of many intel CPUs that made me jump for a yearly upgrade. I mean, SB was great and all, but it really was an exception and not a rule. I've had a lot of CPUs in my time and most of them were used for a 2-3 year period.

 

[Gikaseixas]

Sandy is that good, it makes today's best look average

 

[Pandamonia]

It's not appealing ONLY if you insist on upgrading your CPU every 12 months. On it's own merit, the 4770k is a damn fine chip IMHO, it maintains the exact same IPC increase over it's predecessor as IVB did. I see nothing wrong with that.

I can't think of many intel CPUs that made me jump for a yearly upgrade. I mean, SB was great and all, but it really was an exception and not a rule. I've had a lot of CPUs in my time and most of them were used for a 2-3 year period.

Conroe and Nehalem and Sandy were all great CPU improvements.

 

[frozentundra123456]

Considering the underwhelming reaction elsewhere, you might want to consider that the problem may be with Anand and not Dustin.

The problem was a comparison that was the highest performance A10 vs a ULV intel part.
It is irrelevant want the consensus is on the desktop part. What is even worse is that they didnt even test a standard wattage Ivy Bridge vs Richland. The whole review was like comparing a gas saving economy car to the highest performance model from another manufacturer.

 

[Ajay]

Considering the underwhelming reaction elsewhere, you might want to consider that the problem may be with Anand and not Dustin.

Anand seems to have a great working relationship with Intel. I doubt he would risk that.

As far as I'm concerned, Dustin is right, to an extent anyway. One would think there would be some competition over enthusiast $$s; heck there is competition over cheap smartphones where margins are razor thin.

Sadly, AMD is tied to that albatross GFL that can't seem to meet it's own deadlines on new node transitions (+/-3 yrs), otherwise they could be in a position to fire back with a 20nm XV product in time for Skylake - not as power efficient, but looking at the work done on SR, it could pack a nice jump in IPC. Oh how we miss you now AMD...

 

[Khato]

While I wouldn't have said it in the same way, the quality of the article as a whole simply wasn't up to par and the conclusion made it quite clear why. There are two points about it that really bug me.

1. It's a review of mobile Richland. What business exactly does commentary on desktop Haswell have in the conclusion? I'll agree that it's a perfectly valid complaint - the lack of competition allows Intel to focus its resources on the markets that matter and not have to worry about losing the performance halo. That said, it sure seems like everyone wants to pretend that Haswell would have been completely different from what it is and had a larger performance bump and overclocked better if AMD was still competitive... yeah, that's just wishful thinking.

2. On the actual subject matter of the review he actually does bring up a point that's all too often glossed over - the part being reviewed is the best that AMD has to offer and the lower end parts take pretty healthy hits in GPU performance... But just three paragraphs before he claims that "AMD continues to offer superior mobile graphics" because AMD's top of the line 35W A10-5750M beat out an ultrabook using Intel's 15W i7-4500U with GT2 graphics?!? From previous Anandtech reviews of Iris Pro and the Macbook air he has access to benchmark results for all of the included games (or most at least) running on 15W GT3, 47W GT3e, and 57W GT3e... but instead of including them he simply published with results from the 15W GT2 just to make AMD look better and be able to declare them the winner. How is that claim not a blatant lie?