Seeking Alpha
  • Justin Jaynes
    $AMD - $INTC - Mobile Kaveri ~equal with i5/i7 ULV chips in CPU Perf, crushes them in GPU perf
    Jun 4, 1:40 AM
    Reply (76)
    • Justin Jaynes: GPU perf is about 75% or so of the GTX 750M from Nvidia
      Jun 4, 1:40 AM
    • Justin Jaynes: in Cinebench OpenGL/CPU tests
      Jun 4, 1:41 AM
    • Neosephiroth86: @Justin How about laptop wins? I heard they will announce them only after that bad?
      Jun 4, 2:35 AM
    • Tri Duong: You're comparing a 35W AMD APU vs a 15W Intel CPU. I love the mobile Kaveri but you need to disclose that fact
      Jun 4, 7:30 AM
    • stockalicious: Kaveri lappies look promising - me want one! thats if OEM's dont wrap plastic around it.
      Jun 4, 8:12 AM
    • Tri Duong: There's actually a few AMD laptops that will hit the $600+ range. I think the next gen will get more if AMD lower power another 25% next gen
      Jun 4, 8:21 AM
    • Tri Duong: They really need to work on power consumption and brand awareness/perception.
      Jun 4, 8:27 AM
    • Michael06: It starts with OEMs, and AMD has clearly improved its position. The PRO brand has an entire HP Elite family:
      Jun 4, 8:43 AM
    • Tri Duong: Yea, I'm surprised to see that. HP still playing favorites by naming Intel versions Elite 800s and AMD 700s. At least AMD has a full lineup
      Jun 4, 9:11 AM
    • Justin Jaynes: I don't talk about power consumption because it's pointless to discuss without actual measurements
      Jun 4, 9:18 AM
    • Justin Jaynes: Intel and AMD don't define TDP the same, so 15W Intel vs 35W Kaveri is like saying 15 Apples vs 35W Oranges
      Jun 4, 9:19 AM
    • Justin Jaynes: this chip would align best compared against the 28W Iris Chips from Intel - and here's a graph of a 15W Intel chip consuming 26W under load
      Jun 4, 9:21 AM
    • Justin Jaynes:
      Jun 4, 9:22 AM
    • Tri Duong: We all know the intel chips will win in power consumption. The difference in TDP ratings were off by 10-20%, not 200-300%. Come on man.
      Jun 4, 9:34 AM
    • Justin Jaynes: tri - you've accused me of being biased for weeks - anyone that reads my articles can see I'm not - you even drag private emails into these
      Jun 4, 9:38 AM
    • Justin Jaynes: yes - it's a 35W FX chip compared to an i5 and i7 ULV chip - and I didn't point out power consumption because #1 - I didn't think about it
      Jun 4, 9:40 AM
    • Tri Duong: maybe you don't realize the readers supporting you are also biased towards AMD also? Of course they'll say you're not biased.
      Jun 4, 9:43 AM
    • stockalicious: "such a legend in your own mind" hehe
      Jun 4, 9:43 AM
    • Tri Duong: I've read plenty of your articles. My IT friends read it, my investment bankers friends read it, my gamer friends read it. Same concensus!
      Jun 4, 9:44 AM
    • Justin Jaynes: and #2 - power consumption TDPs between AMD and Intel are only loosely based on reality - look at the time stamp - I threw it up 7 hours ago
      Jun 4, 9:45 AM
    • Justin Jaynes: while I was going to bed after I finished reading the reviews - the fact that AMD is getting within a stone's throw of haswell ULV (15W) CPU
      Jun 4, 9:46 AM
    • Justin Jaynes: perf seems more than you expected - we have no idea what power consumption on the CPU, GPU, or Combined power consumption is - all we have
      Jun 4, 9:48 AM
    • Justin Jaynes: are ratings from AMD and Intel, and a statement from AMD that while web browsing on a 50Whr battery we should get around 9.2 hours of life
      Jun 4, 9:48 AM
    • Justin Jaynes: you have a full 8 CU GPU enabled on the FX 7600P that shouldn't be doing too much work during CPU testing in cinebench - so how much of
      Jun 4, 9:51 AM
    • Justin Jaynes: that 35W TDP will be going idle if the huge GPU isn't really firing?
      Jun 4, 9:51 AM
    • Tri Duong: Yes, it's looking likely that AMD can get their laptop market share back and maybe a little more. HP giving them a full lineup is phenomenal
      Jun 4, 9:51 AM
    • Justin Jaynes: and go try and find some really detailed timed analysis of haswell ULV power consumptions - you'll notice most review sites quit posting
      Jun 4, 9:51 AM
    • Justin Jaynes: rarely post detailed timed power consumption data on haswell ULV. The only one I can find that shows any decent data is the one from AT that
      Jun 4, 9:53 AM
    • Tri Duong: The most important part is their APU moving up to midrange. Hopefully, OEMs will pack respectable batteries, SSD and screens in more configs
      Jun 4, 9:53 AM
    • Justin Jaynes: shows a 15W Intel chip using 26W under load.
      Jun 4, 9:53 AM
    • Justin Jaynes: I'm not biased - I just #1 - didn't put a lot of thought into that post because it was literally typed as I was falling asleep
      Jun 4, 9:54 AM
    • Justin Jaynes: #2 - In your own opinion only you can be right - anyone can look through the condescending nature of our discussions and pick up on this
      Jun 4, 9:55 AM
    • Justin Jaynes: #3 - I didn't think very hard about the power consumption because there isn't any measurements for comparison yet - other than TDP - so
      Jun 4, 9:57 AM
    • Tri Duong: Justin, we both know the 15W chip can run to 30W under turbo then throttle back. Once saturated, it throttles back to normal.
      Jun 4, 9:57 AM
    • Justin Jaynes: I didn't put a lot of thought into it - you're creating a straw man because I didn't point out TDP here - and look at your first post
      Jun 4, 9:58 AM
    • Justin Jaynes: you're unable of discussion without coming across as condescending - you're extremely biased - you think you're the only ethical writer on
      Jun 4, 9:59 AM
    • Justin Jaynes: SA, so you'll go out of your way to try and prove this. You said mobile Kaveri would be extremely lack luster, so now you're pointing out
      Jun 4, 10:00 AM
    • Justin Jaynes: a flaw in my argument, trying to demonstrate how you were right all along - feel free, but saying you're unbiased is a partial truth
      Jun 4, 10:01 AM
    • Tri Duong: That's just how the turbo mode works. Their 1.3ghz turbo to 2.6ghz. AMD's turbo is just very conservative. Overall, Intel consume much less
      Jun 4, 10:03 AM
    • Tri Duong: Just because the chip is consuming 26W doesn't mean it is dissipating 26W. There's a heat sink. It throttles once saturated by the heat sink
      Jun 4, 10:04 AM
    • Tri Duong: Overall, Intel's solution will still consume much less power since turbo mode is short.
      Jun 4, 10:06 AM
    • Justin Jaynes: here's the background: "digihound" is Joel Hruska on ExtremeTech - he helps me understand some of the technical details when I get too far
      Jun 4, 10:08 AM
    • Justin Jaynes: in the weeds - here's his explanation as to why you can't really compare TDPs between Intel and AMD:
      Jun 4, 10:08 AM
    • Justin Jaynes: He's written for PCMag, HotHardware, ARSTechnica, and is now writing for ExtremeTech - and is probably my favorite tech author bar none.
      Jun 4, 10:09 AM
    • Justin Jaynes: I've had very lengthy conversations with Joel about the subject, and have done a lot of digging through benchmarks, and you really can't
      Jun 4, 10:09 AM
    • Tri Duong: I admit I'm biased towards money. I'm never biased towards products. You pump products while I look at business strategy.
      Jun 4, 10:10 AM
    • Justin Jaynes: compare TDPs without looking at actual power consumption - so I usually don't mention power consumption until I have benchmarks.
      Jun 4, 10:11 AM
    • Tri Duong: I'm sorry you feel insulted because I don't think certain superior products may not sell well. This is a financial website after all.
      Jun 4, 10:11 AM
    • Justin Jaynes: Tri is just taking this as an opportunity once again to be condescending and create a straw man to attack. He accuses me of pumping Kaveri
      Jun 4, 10:13 AM
    • Justin Jaynes: because I pointed out it would be better than Richland,but still not catch Intel. This is what I said before:
      Jun 4, 10:17 AM
    • Tri Duong: I own AMD stocks, not Kaveri chips. I only care how AMD will make money. Designing superior architectures is only one part of the equation.
      Jun 4, 10:17 AM
    • Justin Jaynes: tri is doing nothing more right now than trying to create a straw man to prove "he's right" -
      Jun 4, 10:18 AM
    • Tri Duong: You can discuss with Joel all you want but it only matters how useful AMD's products are to the consumer. If they are even aware of AMD...
      Jun 4, 10:18 AM
    • Justin Jaynes: that you'll get no argument on me - product wins matter the most
      Jun 4, 10:19 AM
    • Justin Jaynes: here's my issue with you Tri - you're extremely condescending and don't take criticism well - if you look at the OP - it was a technical
      Jun 4, 10:24 AM
    • Justin Jaynes: discussion - not one about design wins or revenue - if you were less condescending, I would probably have agreed with your 15W vs 35W post
      Jun 4, 10:25 AM
    • Justin Jaynes: but you tell me I "need" to discuss TDP differences and accuse me of being disengenious
      Jun 4, 10:27 AM
    • Justin Jaynes: and at the end, you move the goal posts to design wins and consumers - consumers care about form factor and battery life
      Jun 4, 10:28 AM
    • Tri Duong: lol ok. I guess Intel Atom crushes APUs in power consumption and price. I'll just ignore the performance part. Technical...
      Jun 4, 10:29 AM
    • Justin Jaynes: AMD claims 9.2 hours on a 50 Whr battery while browsing, and that the 35W can squeeze into the same cooling solution as ULV chips
      Jun 4, 10:30 AM
    • Justin Jaynes: "...lower thermal density, the FX-7500 can apparently squeeze into systems with the exact same cooling solutions as the i5-4500U."
      Jun 4, 10:30 AM
    • Justin Jaynes:
      Jun 4, 10:30 AM
    • Tri Duong: Yes, "squeeze" into will result in lower battery life vs intel. That doesn't mean OEMs will even if the APU had better power consumption.
      Jun 4, 10:34 AM
    • kjurden: @Tri Duong: There's plenty of room for opinions on SA, but I can guarantee you if you want to lose viewership/readers...keep this crap up.
      Jun 4, 10:36 AM
    • Tri Duong: Even if HP decides to go all AMD, they can't supply that many chips. They don't have marketing either. Intel got everyone by the balls
      Jun 4, 10:36 AM
    • Justin Jaynes: lower battery life when gaming - 9.2 hours of battery life when browsing is good.
      Jun 4, 10:36 AM
    • Tri Duong: OEMs are struggling, if Intel decides they get less chips then they lose marketshare to their competitors. AMD has low production capacity
      Jun 4, 10:38 AM
    • Tri Duong: Basically, how far AMD can go in the X86 space is largely predetermined by Intel. Servers and consoles are their best bet to break the trend
      Jun 4, 10:39 AM
    • Justin Jaynes: mopeds get better gas mileage than cars, doesn't mean their fun. comparing battery life under gaming is like opting for a moped
      Jun 4, 10:40 AM
    • Justin Jaynes: cooling solution = form factor - and no, I'm not arguing that Intel will still be more efficient - I'm just arguing against this straw man
      Jun 4, 10:41 AM
    • Justin Jaynes: you're creating.
      Jun 4, 10:41 AM
    • Tri Duong: "mopeds get better gas mileage than cars, doesn't mean their fun" 16mp smartphones sell better than 6mp smartphones doesn't mean it's better
      Jun 4, 10:44 AM
    • lostbusker: @Tri Duong, are you picking a fight for the sake of a fight? Take it easy comrade.
      Jun 4, 11:28 AM
    • Michael06: Right, let's come back to power usage once we have data to chew on. TDPs are like unregulated mileage estimates for cars.
      Jun 4, 11:37 AM
    • Michael06: Kaveri offers a nice value proposition and should improve or at least sustain laptop share. As investors, we can all celebrate this.
      Jun 4, 11:37 AM
    • inside man 55: @Michael agree, good to see AMD have so many partners to launch mobile Kaveri, this is a good sign for back to school sales.
      Jun 4, 11:46 AM