<?xml version="1.0" encoding="utf-8"?>
<rss version="2.0" xmlns:dc="http://purl.org/dc/elements/1.1/" xmlns:atom="http://www.w3.org/2005/Atom">
	<channel>
		<title>Intel redesigns the transistor</title>
		<description>Discuss Intel redesigns the transistor</description>
		<link>http://www.fudzilla.com/home/item/22642-intel-redesigns-the-transistor</link>
		<lastBuildDate>Thu, 20 Jun 2013 07:23:41 +0100</lastBuildDate>
		<generator>JComments</generator>
		<atom:link href="http://www.fudzilla.com/component/jcomments/feed/com_k2/22642/10" rel="self" type="application/rss+xml" />
		<item>
			<title>weasel says:</title>
			<link>http://www.fudzilla.com/home/item/22642-intel-redesigns-the-transistor#comment-23579</link>
			<description><![CDATA[X64 added more registers not SSE2, and it was only an extra 8. The cache control instructions are next to useless in all but a few very specific cases. And no one actually uses them IME. You could already do horizontal arithmetic with existing instructions, it was just a PITA. They're convenient, but they don't do anything special. They aren't even much faster, only about 20%. It's all irrelevant. If I went to my boss and went into "detail" about what special instructions I'd used in my dsp code he'd say "STFU and show me the benchmarks".]]></description>
			<dc:creator>weasel</dc:creator>
			<pubDate>Fri, 27 May 2011 21:54:07 +0100</pubDate>
			<guid>http://www.fudzilla.com/home/item/22642-intel-redesigns-the-transistor#comment-23579</guid>
		</item>
		<item>
			<title>Naterm says:</title>
			<link>http://www.fudzilla.com/home/item/22642-intel-redesigns-the-transistor#comment-22635</link>
			<description><![CDATA[You posted a SINGLE synthetic benchmark and you act as though you wrote a [censored]ing white paper. There simply aren't many thorough comparisons of the two processors online because they operate in differing market segments. Intel is trying to shoe horn x86 into the mobile space, ARM is trying to build itself up to hyperscale servers and more traditional mobile computing markets. For the moment, they're not really competing. Maybe in another 2-3 years. The only thorough comparisons of ARM and Atom I can find are done with A8. It was a lot of A8 choking badly on heavy floating point maths and javascript. It was a hyperscale server feasibility test which the ARM processors failed. Obviously, those would be irrelevant.]]></description>
			<dc:creator>Naterm</dc:creator>
			<pubDate>Thu, 12 May 2011 09:47:59 +0100</pubDate>
			<guid>http://www.fudzilla.com/home/item/22642-intel-redesigns-the-transistor#comment-22635</guid>
		</item>
		<item>
			<title>Naterm says:</title>
			<link>http://www.fudzilla.com/home/item/22642-intel-redesigns-the-transistor#comment-22634</link>
			<description><![CDATA[Why do you keep propping up these straw men? I'm well aware of ARM ISA processor's lower power consumption. SSE/2/3/SSSE3/4.x are SIMD extensions for x86. SSE added eight, 128-bit registers. SSE2 added 32 new registers onto that. It's other claim to fame is the implementation of cache control instructions. SSE3 allowed horizontal workflow in a register while also allowing the more efficient conversion of floating point values to integer. SSSE3 added 16 more instruction and 47 more with 4.1 and a further 7 with .2. That's really about as much detail as I'm going to get into in 700 words. I'm well aware of NEON.]]></description>
			<dc:creator>Naterm</dc:creator>
			<pubDate>Thu, 12 May 2011 09:43:49 +0100</pubDate>
			<guid>http://www.fudzilla.com/home/item/22642-intel-redesigns-the-transistor#comment-22634</guid>
		</item>
		<item>
			<title>JonhyW says:</title>
			<link>http://www.fudzilla.com/home/item/22642-intel-redesigns-the-transistor#comment-22629</link>
			<description><![CDATA[Think in terms of power consumption. ARM architecture is more than 4x power efficient than x86 for same performance. You don´t talk about power terms because you know that Intel has nothing to propose for tablets and smartphones. Atom runs too hot and is power-hungry. I don´t see any company excited about Atom.Samsung, Apple, Nvidia, Qualcomm, LG, Texas Instruments, Microsoft are investing millions in ARM architecture. AMD and ARM are discussing the possibility of business together. It´s silly to wait x86 be successful for mobile. Intel is strong where it is. But some people really exaggerate and think that Intel can do magic and will beat every GPU with its Larabee. Seriously? And now Intel fans believe that Atom will beat ARM?]]></description>
			<dc:creator>JonhyW</dc:creator>
			<pubDate>Thu, 12 May 2011 04:17:52 +0100</pubDate>
			<guid>http://www.fudzilla.com/home/item/22642-intel-redesigns-the-transistor#comment-22629</guid>
		</item>
		<item>
			<title>JonhyW says:</title>
			<link>http://www.fudzilla.com/home/item/22642-intel-redesigns-the-transistor#comment-22627</link>
			<description><![CDATA[You haven´t provide any valid information, compatison,link s. Nothing. Just another bunch of words. Please, be serious next time. You don´t have idea about what is SSE and x86 and ARM equivalent of it. Your poor wording doesn´t work. SSE? Do you know what are you talking about, boy?]]></description>
			<dc:creator>JonhyW</dc:creator>
			<pubDate>Thu, 12 May 2011 04:00:34 +0100</pubDate>
			<guid>http://www.fudzilla.com/home/item/22642-intel-redesigns-the-transistor#comment-22627</guid>
		</item>
		<item>
			<title>Naterm says:</title>
			<link>http://www.fudzilla.com/home/item/22642-intel-redesigns-the-transistor#comment-22540</link>
			<description><![CDATA[I did. A quick check of Google shows a lot of dual core A9s beating up on single core, 1.6GHz Atoms. Too bad it's not 2008. There aren't a lot of good, traditional benchmarks running serious, SSE supported software on an A9 because the two processors exist in different ecosystems. You'd wager an A9 would be faster than an Atom doing an MPEG2 to H.264 transcode with software with SSE support? I sure as hell wouldn't. Run some heavy floating point on both of them. Run some heavy Javascript on both of them. All you have is one single synthetic benchmark. That's not nearly enough. On a side note, you need to calm down a little bit buddy. I know you're all of 13 years old, but try to act like a grown-up.]]></description>
			<dc:creator>Naterm</dc:creator>
			<pubDate>Wed, 11 May 2011 09:06:14 +0100</pubDate>
			<guid>http://www.fudzilla.com/home/item/22642-intel-redesigns-the-transistor#comment-22540</guid>
		</item>
		<item>
			<title>JonhyW says:</title>
			<link>http://www.fudzilla.com/home/item/22642-intel-redesigns-the-transistor#comment-22439</link>
			<description><![CDATA[Do you know what are you talking about? All we see it is you just talking fairy tails. Just words. Where is the tests, benchmarks, description? Stop thinking that people will beleive you just because you will say it. It is naive and childish.]]></description>
			<dc:creator>JonhyW</dc:creator>
			<pubDate>Tue, 10 May 2011 04:56:03 +0100</pubDate>
			<guid>http://www.fudzilla.com/home/item/22642-intel-redesigns-the-transistor#comment-22439</guid>
		</item>
		<item>
			<title>Naterm says:</title>
			<link>http://www.fudzilla.com/home/item/22642-intel-redesigns-the-transistor#comment-22373</link>
			<description><![CDATA[Regardless, two A9 cores are roughly equivalent to one Atom core. SMT isn't an extra core, it's maybe a 10% boost. The point still stands, Atom cores are more powerful than A9 cores. Run something that has some serious SSE optimizations, see how well they compare then.]]></description>
			<dc:creator>Naterm</dc:creator>
			<pubDate>Mon, 09 May 2011 12:20:00 +0100</pubDate>
			<guid>http://www.fudzilla.com/home/item/22642-intel-redesigns-the-transistor#comment-22373</guid>
		</item>
		<item>
			<title>JonhyW says:</title>
			<link>http://www.fudzilla.com/home/item/22642-intel-redesigns-the-transistor#comment-22335</link>
			<description><![CDATA[Nice joke. There weren't four cores. http://2.bp.blogspot.com/_pEfSMUbljUA/S9yFTQn2xoI/AAAAAAAAAD8/P2r0WkVYhZI/s1600/Benchmark_ipad_iphone_atom_arm_cortex_apple_a4_nvidia.JPG Performance: Dual core ARM Cortex A-9 (2x1 GHz) is faster then Atom (1core/2threads 2GHz). TDP: Cortex A-9 Dual core +GPU - 0.60-0.75 Watts. Atom N550 - 8.5 Watts. Atom N450 - 5.5 Watts. Atom N280 - 2.5 Watts (but no GPU) Atom Oak Trail - Z670 - still 3 Watts. No words. Add to it that Atom is much expensive.]]></description>
			<dc:creator>JonhyW</dc:creator>
			<pubDate>Sat, 07 May 2011 05:45:49 +0100</pubDate>
			<guid>http://www.fudzilla.com/home/item/22642-intel-redesigns-the-transistor#comment-22335</guid>
		</item>
		<item>
			<title>Naterm says:</title>
			<link>http://www.fudzilla.com/home/item/22642-intel-redesigns-the-transistor#comment-22334</link>
			<description><![CDATA[You just answered your own question. Four A9 cores are roughly as fast as one Atom core. I'd say that pretty much sums it up. Maybe you need to read a post, really read it, before you post unnecessary responses. I see why intel is going after this market, but I don't see why they're doing it with x86. If they're worried about not selling enough x86 chips, they shouldn't be. All these new client devices require massive server resources to work properly. The server market is basically x86 now with intel dominating from a performance and power perspective. I think it's unlikely that AMD will do anything to change that in the near term, if ever. I don't know why intel sold off all their ARM assets, they'd be pretty useful now.]]></description>
			<dc:creator>Naterm</dc:creator>
			<pubDate>Sat, 07 May 2011 05:19:10 +0100</pubDate>
			<guid>http://www.fudzilla.com/home/item/22642-intel-redesigns-the-transistor#comment-22334</guid>
		</item>
	</channel>
</rss>
