Qualcomm cuts 4,500 jobs, may spin off divisions

[Thala]

It's possible that the physical layout of their Cortex-A57 for TSMC 20nm is sub-par compared to Exynos 7420's.

Physical layout is Qualcomms job, as Cortex-A57 is delivered as synthesizable IP. In addition Samsung had no issues with A57 on 20nm. In conclusion it is all Qualcomms fault.

 

[mrmt]

Even Intel does not have that kind of money. Qualcomm's Market cap is a huge $100B. It's almost as big as Intel which has a Market cap of $136B.

The bulk of the value is due to the licensing business, not because of the chip business, so the actual number for an acquisition is much, much smaller. The other issue is that Intel could engineer a merger, all-cash acquisitions aren't what we should expect from a deal of that size.

But in any case, the fact that Qualcomm is thinking about spinning off its chip business says a lot of what they expect in terms of quality and profitability of what they have in the pipeline.

 

[el etro]

Ah good, the Samsung FUD machine is working at full speed.

Whoa! Then it definitively all this was not an attack to the ARM A57 core(a core that really have problems, i know).

 

[Exophase]

Physical layout is Qualcomms job, as Cortex-A57 is delivered as synthesizable IP. In addition Samsung had no issues with A57 on 20nm. In conclusion it is all Qualcomms fault.

ARM, in partnership with third parties and the fabs, also sells pre-made physical macros, which the smaller SoC makers have been using for years.

Recently they've been promoting that they've stepped up their efforts to make these macros more competitive with the custom work from top-tier SoC makers but they're probably not there yet, particularly not with the 20nm TSMC part and not in the timeframe Qualcomm needed it.

But I wouldn't discount the possibility that Qualcomm did in fact use a sub-par ARM physical macro as opposed to using a rushed internally designed layout. There are rumors that Qualcomm outsourced much of the CPU design in S810 which could have amounted to this.

 

[2is]

Yeah but for a phone that's alright.

Maybe if everyone was using your chip, but when the reviews are out and you're slower than everyone else, it's not alright, and neither is an overheating device and that's why they're in this hole.

 

[monstercameron]

Maybe if everyone was using your chip, but when the reviews are out and you're slower than everyone else, it's not alright, and neither is an overheating device and that's why they're in this hole.

Who's buys phones with benchmarks?

 

[erunion]

Oh boy... The way they are falling down is way faster than AMD!

Not even comparable.

Qualcomm is making hard cuts early, rather than doubling down on a failed strategy.

Qualcomm has a profitable business it can fall back on, both of AMDs businesses are in decline.

 

[2is]

Who's buys phones with benchmarks?

You're either talking before thinking or you've lived under a rock for the past half decade or more, otherwise you'd have noticed that dozens upon dozens of tech sites now review phones and benchmarks are a key talking point in these reviews. Why? Because people read them, the same people who buy phones, and you can be pretty sure that these sites don't waste their time running benchmarks if consumers didn't care about it and skipped that part of the review. You can also bet companies like Samsung and Qualcomm among others wouldn't spend hundreds of millions in R&D costs to make faster processors for phones if nobody cared.

 

[Exophase]

You're either talking before thinking or you've lived under a rock for the past half decade or more, otherwise you'd have noticed that dozens upon dozens of tech sites now review phones and benchmarks are a key talking point in these reviews. Why? Because people read them, the same people who buy phones, and you can be pretty sure that these sites don't waste their time running benchmarks if consumers didn't care about it and skipped that part of the review. You can also bet companies like Samsung and Qualcomm among others wouldn't spend hundreds of millions in R&D costs to make faster processors for phones.

Yeah, and the people who read the reviews generate a lot of buzz for the people who don't. They get asked for advice on purchases and this ends up factoring into it and it all spreads from there. Just planting any kind of seed of perceived superiority can go a long way.

 

[monstercameron]

You're either talking before thinking or you've lived under a rock for the past half decade or more, otherwise you'd have noticed that dozens upon dozens of tech sites now review phones and benchmarks are a key talking point in these reviews. Why? Because people read them, the same people who buy phones, and you can be pretty sure that these sites don't waste their time running benchmarks if consumers didn't care about it and skipped that part of the review. You can also bet companies like Samsung and Qualcomm among others wouldn't spend hundreds of millions in R&D costs to make faster processors for phones if nobody cared.

I agree that benchmarks are important but how do you explain the iphone? definitely not a benchmark beater yet very popular. What about samsung, even when they arent the fastest they still sell well. The normies/non-techies dont buy benchmarks.

 

[Idontcare]

I agree that benchmarks are important but how do you explain the iphone? definitely not a benchmark beater yet very popular. What about samsung, even when they arent the fastest they still sell well. The normies/non-techies dont buy benchmarks.

When it comes to the iPhone the explanation is in the apps. iTunes/iApps across both iPhone and iPad makes the full package experience.

Mobile benchmarking misses the demographic entirely because it assumes the demographic cares about performance. Really it just cares about stuff working together, working acceptably well, and being cheaply accessible.

Nail that and your rewards are billions in profit every 90 days.

 

[monstercameron]

When it comes to the iPhone the explanation is in the apps. iTunes/iApps across both iPhone and iPad makes the full package experience.

Mobile benchmarking misses the demographic entirely because it assumes the demographic cares about performance. Really it just cares about stuff working together, working acceptably well, and being cheaply accessible.

Nail that and your rewards are billions in profit every 90 days.

I agree, too many variables for benchmarks to be the most weighted category. Even for sammy, features that we forum elite might scoff at as gimmicky, that probably sells more units than winning a geekbench or the like.

Point being performance is important but experience is the most important.

 

[2is]

I agree that benchmarks are important but how do you explain the iphone? definitely not a benchmark beater yet very popular. What about samsung, even when they arent the fastest they still sell well. The normies/non-techies dont buy benchmarks.

Apple also has end to end control which includes iOS optimization and a custom designed SoC which may not win in a numbers game due to it being a dual core while many have moved to quads, but in actual performance the iPhone is just as fast as any other phone. It's an exception, not the rule. Most Android phones have to brute force their way to a smooth UI and still can't get there except for maybe the S6, which Samsung has coincidentally is also decided to develop a custom SoC for.

 

[Phynaz]

I agree that benchmarks are important but how do you explain the iphone? definitely not a benchmark beater yet very popular. What about samsung, even when they arent the fastest they still sell well. The normies/non-techies dont buy benchmarks.

Typically an idevice is the fastest in its product class at release.

 

[Arachnotronic]

Apple also has end to end control which includes iOS optimization and a custom designed SoC which may not win in a numbers game due to it being a dual core while many have moved to quads, but in actual performance the iPhone is just as fast as any other phone. It's an exception, not the rule. Most Android phones have to brute force their way to a smooth UI and still can't get there except for maybe the S6, which Samsung has coincidentally is also decided to develop a custom SoC for.

I have both an S6 and an iPhone 6, and the UI is still meaningfully smoother on the iPhone 6.

Apple is, IMO, untouchable when it comes to a fast and fluid smartphone experience.

 

[Idontcare]

All set to buy AMD!

Call me cynical, but leadership management in crisis mode is a dangerous quarry, steady hands yesterday become capricious hands today, an AMD M&A is infinitesimally more likely today than it was yesteryear.

 

[Phynaz]

IMHO Qualcomm got what they wanted from AMD in 2009.

 

[Exophase]

I agree that benchmarks are important but how do you explain the iphone? definitely not a benchmark beater yet very popular. What about samsung, even when they arent the fastest they still sell well. The normies/non-techies dont buy benchmarks.

iPhone does extremely well both in CPU and GPU performance, including many benchmarks. It makes a compromise in having two cores over 4+, which is not a one-sided performance loss since it allows them to devote more resources to single threaded performance. Yet even then, it still manages to top a lot of threaded benchmarks.

Samsung is also highly competitive. So I don't think you've picked the two best examples.

Even if Apple weren't very well positioned in performance they have their own highly valued OS and app ecosystem so they could play under different parameters, but this doesn't apply to anyone else.

 

[cbn]

Problem is when companies do this they shot themselves in the foot. 1 division may have a hard time an held up by another but then that other has issues then they may be held up by the other.

What happens when the old patents expire and its the CPU division that gets a good run or comes up with new patents. Seems these 2 can feed off each other and would be better suited together vs apart. But that is long term, these people want a quick buck so they do not think long term and that has killed many business's.

When you wrote that the first thing I thought of was Intel's mobile division (ie, the phone SoCs).

Even though Intel's phone SoCs are losing money, I do hope they keep it going. This for the long term sake of the company, rather than for the short term.

 

[ninaholic37]

How are smartphone sales doing in general this year? Is the market and demand dying down? I never liked smartphones or touchscreen anything at all, so overheating chips is a blessing to me! DIE DIE DIE and bring me back the netbook with real full-sized keys instead

 

[Arachnotronic]

How are smartphone sales doing in general this year? Is the market and demand dying down? I never liked smartphones or touchscreen anything at all, so overheating chips is a blessing to me! DIE DIE DIE and bring me back the netbook with real full-sized keys instead

There are tons of cheap Bay Trail notebooks that you can buy if you've got a thing for Netbooks...

 

[Ancalagon44]

Let me get this straight.

When the company makes a engineering/product mistake, the right way to correct that is to fire/retrench the people with the knowledge and experience to correct the mistake?

 

[BUnit1701]

Let me get this straight.

When the company makes a engineering/product mistake, the right way to correct that is to fire/retrench the people with the knowledge and experience to correct the mistake?

Of course, have to drive 'shareholder value' above all other considerations, even longevity or R&D.

 

[Ancalagon44]

Of course, have to drive 'shareholder value' above all other considerations, even longevity or R&D.

If the shareholders had any sense, they would sell after an announcement like this.

 

[Tsavo]

Let me get this straight.

When the company makes a engineering/product mistake, the right way to correct that is to fire/retrench the people with the knowledge and experience to correct the mistake?

It worked for AMD!