分析师：AMD在ARM身上看不到未来

ARM近段时间来一直在积极游说英特尔的长期对手AMD，促使其放弃x86架构，成为其授权客户。 这对于AMD来说毫无意义，至少有一名分析师持这种看法。Raymond James & Associates Inc.的分析师Hans Mosesmann在一份报告中写道：“AMD坚持x86，目前就是这样。” “毫无疑问，除非AMD换新的CEO，否则他们是不会考虑采取ARM处理器内核策略的，因为这些都是次要的考量，”他说，“我们的浅见是，AMD进入ARM市场不会给其带来附加价值。” 今年稍早的时候，AMD总裁兼首席执行官Dirk Meyer在他上任两年后辞职。其职位由AMD首席财务官兼高级副总裁Thomas Seifert暂代。当时有报导暗示，他是被迫辞职，因为董事会对AMD缺乏主打平板电脑和移动市场的芯片感到不满。（详见文章：评论：为什么ARM与AMD合作有意义？） Mosesmann在报告中表示，他在机缘巧合下曾经“偶遇”AMD高级副总裁兼首席营销官Nigel Dessau并问到了一些东西。他说：“Dessau改变了AMD以往缺乏平板方案的策略，但称AMD暂时还只是把它当作超低功耗市场的一部分。因为考虑到平板电脑现在仍然不成气候，AMD打算在其Fusion系列低功耗产品常态缩减计划中，追求更大量的平板设备市场。” AMD可能在另一个方面颠覆市场。“今年稍晚，Bulldozer 即将推出，对AMD来说可能诱惑更大的是服务器云/数据中心领域，”Mosesmann表示，“我们还不确定Bulldozer是否能显著改变企业应用领域的竞争态势，但值得庆贺的是Opteron品牌终于后继有人，而且可能在大型数据中心应用中有出色表现。在这类应用中，“性能/功率/成本”的考虑，可能比以往单纯的“性能”或“性能/功率”的考虑更优秀。” 2010年，AMD首次公开介绍Bulldozer：一款从主流笔记本到高端服务器的所有产品通吃的核心，将在2011年底之前上市。该产品上市拖了较长时间，在一定程度上因为它是GlobalFoundries的32纳米工艺生产的第二款芯片。 GlobalFoundries生产的第一款主要32纳米处理器是AMD以前宣布的Llano，它是AMD的Fusion系列芯片中的第一款产品。它也是AMD第一款基于高k和金属栅(high-k /metal gates)的处理器。Fusion系列把x86与图形内核集成到了一起。 “Llano ，或着叫APU(CPU和图形内核集成在一个芯片上)在6月(可能是在台北国际电脑展Computex期间) 发布的时候，期望AMD能在消费PC领域得到大量订单，”该分析师表示，“Llano本身未必是AMD能否在商业PC和或/服务器市场中取得成功的关键，但它是一种急需的解决方案，可以填补AMD产品组合中的一个空白。” Llano适合哪里？“Llano的定位，将大致是在英特尔基于Sandy Bridge的酷睿i3所处的领域，也许是酷睿i5的低端领域，”他说，“Llano的定位似乎略低于或等于我们先前的预期。但是，考虑到低端Brazos Fusion APU(面向上网本那样的价格点)所得到的成功，我们猜测市场的接受情况会非常良好。谈到英特尔强调的Sandy Bridge图形性能，AMD笑而不语，因为Llano在传统图形指标方面(classical graphics metrics)应该能够轻松击败Sandy Bridge。” 本文授权编译自EE Times，版权所有，谢绝转载 参考英文原文：Analyst: AMD-ARM deal makes no sense, by Mark LaPedus

如果今天你是AMD的新CEO，那么你会：
	1. 安全第一，继续当英特尔的追随者		2. 转型！做基于ARM的领跑者
	3. 让平板飞一会，再决定下一步		4. 以前的架构都太逊了，我要创造新世界

c3uesmc

相关阅读：
• 评论：为什么ARM与AMD合作有意义？
• ARM还是x86：GPU大佬们移动领域的豪赌
• 英特尔赢得三栅极晶体管竞赛短程胜利，独缺移动处理器
• 重用自家芯，2010苹果产品A4用量4倍于X86c3uesmc

{pagination} Analyst: AMD-ARM deal makes no sense Mark LaPedus SAN JOSE, Calif. - ARM Holdings plc is trying to persuade Advanced Micro Devices Inc., the long-time rival to global chip leader Intel Corp., to license ARM processors and use them instead of the x86 architecture. This makes little or no sense for AMD-at least according to one analyst. ''AMD is committed to x86; for now,'' said Hans Mosesmann, an analyst with Raymond James & Associates Inc., in a report. ''Not surprisingly until AMD hires a new CEO the big strategic decisions such as pursuing or not an ARM processor core strategy will be on the backburner,'' he said. ''Our two cents: we do not see an AMD value-add in the ARM market.'' Earlier this year, Dirk Meyer, president and CEO of AMD, resigned from the company after two years as its top executive. Thomas Seifert, AMD's chief financial officer and senior vice president, was appointed interim CEO of AMD. In the report, Mosesmann said he recently had a chance to ''catch up'' with AMD senior vice president and chief marketing officer Nigel Dessau. ''Dessau deflected AMD’s lack of tablet strategy other than to say that the company views it simply as an ultra low power part of the market and as such, given tablets are such a small market today, AMD intends to intersect higher tablet volumes as part of the normal scale down of its Fusion portfolio,'' he said. AMD could shake up the market on another front. ''A potentially more interesting catalyst for AMD later this year is in the server cloud/datacenter arena with the upcoming Bulldozer core,'' Mosesmann said. ''We doubt Bulldozer will disrupt much the competitive dynamic in enterprise but it is a long awaited update for the Opteron brand and may play very nicely in large datacenter applications where performance/watt/dollar may trump performance or performance/watt.'' Last year, AMD made its first public presentations on Bulldozer. Bulldozer, which targets everything from mainstream notebooks to high-end servers, will come toward the end of 2011. The longer time to market is due in part to the fact the chip is the second in line to be made in GlobalFoundries' 32-nm process. The fab's first major 32-nm processor is AMD's previously-announced Llano, the first member of AMD's Fusion family which merges x86 and graphics cores. It is also AMD's first processor, based on high-k and metal gates. ''Look for AMD’s consumer PC campaign to get an important volume kick with the launch of the Llano processor, or APU (CPU and graphics in one chip), in June (likely at the Computex show in Taiwan),'' according to the analyst. ''Llano on its own doesn’t solve or relate necessarily to success in commercial PCs and or/the server markets; however, it’s a much needed solution that has been a hole in AMD’s product portfolio.'' Where does Llano fit? ''Llano will be positioned roughly in the segment where the Intel Sandy Bridge-based Core i3 resides and perhaps the lower end of the Core i5,'' he said. ''Llano seems positioned at or slightly below where we had expected; however, given the success of the lower-end Brazos Fusion APU (targeting netbook level price points), we suspect market adoption will be quite reasonable. AMD is delighted in Intel emphasis of Sandy Bridge’s graphics performance as Llano should easily trump it in classical graphics metrics.''