Nvidia’s acquisition of Ageia in 2008 was a strategic move to boost the marketability of its GPU offerings. With the discontinuation of the dedicated PhyX boards, the acceleration moved to the GeForce GPU as a differentiation factor that set it apart from AMD’s ATI cards.
If a PhysX game detected the presence of an Nvidia GPU, it would move the hardware physics to the video card. Without an Nvidia board, the physics would hit the CPU, which in all cases is slower than what a GPU can do.
It’s expected that Nvidia would like to do everything it can to distance itself from the CPU and the GPUs of its competitors, but closer looks at the PhysX software implementation have shown that there could be some shadiness going on.
An excellent investigation by David Kanter at Real World Technologies found that Nvidia’s PhysX software implementation for use by CPUs still uses x87 code, which has been deprecated by Intel in 2005 and now has been fully replaced by SSE. Intel supported SSE since 2000, and AMD implemented it in 2003.
The x87 code is slow, ugly, and remains supported on today’s modern CPU solely for legacy reasons. In short, there is no technical reason for Nvidia to continue running PhysX on CPUs using such terrible software when moving to SSE would speed things considerably – unless that would make the GeForce GPGPU look less mighty compared to the CPU.
Archive
Tag: CPU
Task Manager Free 4.9.0 monitors processes, services, connections, CPU, disk usage, memory, DLLs, drivers. View description of startup programs and services. Get alerts on new startups.
Advanced startup manager allows you to:
# Get full list of applications running automatically on Windows startup including all hidden applications.
# Get all additional information about startup applications.
# Disable/enable, add/delete, edit startup entries.
# Stop or run once again startup application.
Startup guard allows you to:
# See alert when any new application adds itself to startup.
# Forbid some applications to be added to startup.
# See alert when Internet Explorer home page was changed.
Process manager allows you to:
# Get full information about all processes: CPU and memory usage, executable name with full path, priority, work time, user name.
# Get executable file properties, icon and version information.
# Stop many processes with one click.
# Change process priorities.
Tray icons allows you to:
# Get current information about CPU usage including list of most active programs.
# Get current information about HDD usage.
Protection against viruses includes:
# Detection and destruction of most propagated viruses.
# Virus database update.
# Minimum usage of system resources.
Much of the focus was on Harpertown, a 45nm quad-core version of Xeon with 12MB of cache, running in conjunction with Stokley, which support a 1600 MHz front side bus. Intel will also be releasing Wolfdale, a dual-core version with 6MB of cache. Although Penryn is primary a shrink of the existing architecture, rather than a complete new architecture, the new chips do s have some new features, notably including more cache and support for the faster bus; as well as a new divider that is supposed to be faster, and new SSE4 instructions…
The Tech Report compared a 3GHz Harpertown Xeon E5472 (expected to be released Nov. 12) against a 2.5GHz AMD “Barcelona” Opteron 2360SE (expected by the end of the year) and found that the Xeon pretty much won all the tests. The Xeon beat the Operton by 4% at Specjbb (a server business logic benchmark), the new Xeon beat an older 3GHz Clovertown Xeon by about 10%; and the 2.5GHz AMD part by about 4%. On almost all the other tests, Harpertown does even better, ahead of Barcelona by 20 to 34% in real world applications (and more in some synthetic benchmarks I’m more skeptical about).
Anandtech has even more benchmarks, most showing Intel ahead by 27% to as much as 60% (though I’m skeptical of the later tests). But it shows AMD ahead on performance/watt (with a larger improvement the less work is being done), almost certainly because of the more efficient memory scheme.
Both Barcelona and Harpertown seem to have some headroom in clock speeds. AMD is only shipping 2GHz now, but has promised 2.5GHz for December; let’s hope it can eventually match the 3GHz clock that Harpertown will start with. Intel is promising a 3.16GHz Harpertown; and on the desktop side showed a a demo of a 5.5GHz quad-core desktop with a ridiculous amount of cooling.
Looks like AMD has not yet caught back up but we’ll keep an eye out for more quad-core news and comparisons. I certainly hope AMD catches up and not necessarily because I am a fan. AMD has been good to me for quite a while but I don’t hold any loyalties between the two. I like competition. It’s always better for the consumer as, with healthy competition, it pushes the competitors to outdo each other in performance as well as price – two nice things as a consumer I appreciate. Competition can typically make a product better, faster, as well.
System builders who received samples a week or two ahead of today’s worldwide launch say they aren’t ready to issue benchmarks just yet. Nevertheless, sources tell ChannelWeb that the processor AMD calls “the first native quad-core” is faster than they had anticipated. They say three key advances are testing out as advertised — a tri-level memory cache hierarchy with fully shared L3 cache for all four cores, a floating point unit with 2×128-bit loads/cycle, and independent power supplies for each of the processor’s four cores and to the memory controller. The last feature distinguishes AMD’s quad-core product from Intel’s, in that it’s possible to idle one, two or three CPU cores for a workload to better manage power consumption.
CMP Channel’s Test Center last week received an engineering sample server equipped with dual Barcelona CPUs. After putting it through its paces, Frank Ohlhorst reports, “Those who have waited for the arrival of AMD’s next generation CPU won’t be disappointed.”
As far as pricing, AMD is remaining tight-lipped about how it plans to scale its new quad-cores against Intel’s or its own dual-core chips. Partners in the know say Barcelona will be “competitively priced.” Market watchers say it will have to be, given Intel’s recent slashing of its own quad-core prices down to levels nearly in line with its Core 2 Duo products.
If you have set aside a cool US$1100 for your processor alone at the time of this review release, then you should take a look at the QX6850 I will be reviewing today.
With four 8MB cache supported independent processing on-die units blistering along at 3.0Ghz, this latest Core 2 Extreme processor release from Intel at 1333Mhz front side bus sets the bar of what a consumer available processor can reach at this time. The QX6850 is currently holding the performance belt and by what is seen developing in the market from the competition, the only thing that will be beating it will be the next quad-core release from Intel (which will be within about 6 months).
The first Barcelona models, formally called Quad-Core Opteron, will run at clock frequencies up to 2GHz and will be available in standard and low-power versions. Faster models, both of the standard and more power-hungry special-edition ilk, will arrive in the fourth quarter, the company said. The first servers using the chips will come in September…
AMD’s Barcelona puts four cores on a single slice of silicon, an approach AMD calls “native quad-core,” and the company has argued that Barcelona will outperform the Xeon 5300. The only problem: that comparison soon will become obsolete.
Intel’s second-generation quad-core server processors, “Harpertown” a server member of Intel’s “Penryn” family, will arrive this year, too, with the promise of better performance, lower power consumption and lower manufacturing costs by virtue of a manufacturing process with 45-nanometer features. AMD is only just now moving to a 65-nanometer process.
For decades, typical computer processors had a single processing engine, but dual-core models with two engines began arriving this decade as a way to try to improve performance without consuming inordinate amounts of power and producing corresponding amounts of waste heat. Now chipmakers have moved to quad-core and octo-core models; Sun Microsystems plans to debut its 16-core “Rock” chip in 2008.
Putting multiple cores on a chip isn’t a miracle cure, though. For one thing, it’s hard to adapt software for the chips–especially software for PCs.
For another, a chip with four cores consumes more power than an otherwise comparable model with two, so multicore chips typically run at lower clock frequencies to keep power consumption down. Current dual-core Opteron chips run as fast as 2.8GHz.
A faster clock frequency will let a processor execute a given task more quickly, but multiple cores will let it do more jobs at once.
Also this year, AMD plans to release a quad-core chip for PCs. It and high-end dual-core models will sport a new “Phenom” brand.
Intel Releases Quad-Core Chips
Intel Corp. has unveiled two new microprocessor with four cores aimed at embedded market, marking 30 years of its years of delivering products and technology to the embedded computing market. The new processors will enable performance of a multi-processor server with embedded machines.
The new processors for embedded systems are quad-core Intel Xeon E5335 and E5345, which are already available for LGA771 infrastructure. The chips operate at 2.0GHz and 2.33GHz, respectively, use 1333MHz processor system bus and contain 8MB of level-two cache. Thermal design power (TDP) of the new central processing units was not announced, however, Intel Xeon models E5335 and E5345 in LGA771 had TDP of 80W.
The quad-core Intel Xeon processors E5335 and E5345 are available today starting at $690.
In addition to the new processors, Intel also unveiled IP Network Server NSC2U which supports quad-core Xeon processors and suits a variety of network-centric applications, from security intrusion prevention to telecommunications services-over-IP (SoIP), including IMS, IPTV and Video on Demand (VoD). The NSC2U server features a ruggedized chassis, compact form factor and extended lifecycle support for the components. The Intel IP Network Server NSC2U is targeted for July 2007 availability.
The launch of the new quad-core chips for embedded systems commemorates 30 years of Intel’s presence in embedded computing market. In 1976, before the advent of the personal computer, Intel entered the embedded market and provided extended life cycle support for technologies, silicon and platforms that drive this industry. Intel embedded technology can be found today in automobiles, airplanes, ATMs, information kiosks, telecommunications infrastructure and network storage systems, as well as factory and medical equipment.
AMD Improves Price-Performance Ratio of Its Chips
Advanced Micro Devices on Monday officially reduced prices of its desktop microprocessors, making dual-core chips available at less than $100 price-points and reducing the price of its top-of-the-range offerings from $999 to $799. The price-cut may boost interest towards AMD’s microprocessors, but the fierce price war with Intel reduces both chipmakers’ profits.
The furious competition from Intel’s Extreme processor series has forced AMD to start selling two Athlon 64 FX microprocessors for $999, the price-point of one former top-of-the-range offering, and now the firm had to reduce the price-tag on its fastest desktop solution that consists of two AMD Athlon 64 FX-74 chips by 20% to $799. In addition, AMD removed AMD Athlon 64 FX-70 and FX-62 central processing units from its price-list, leaving only two processor bundles – FX-72 and FX-74 – in the family.
AMD’s dual-core Athlon 64 X2 family also faced significant price slashes from 23% to 48%, making the fastest desktop processor from the company – AMD Athlon 64 X2 6000+ – to cost $241. The least expensive dual-core chips from AMD – Athlon 64 X2 4000+, X2 3800+ and X2 3600+ – now cost $104, $83 and $72 in 1000-unit quantities, making two processing engines from AMD available in value segment. continue reading…
Few computing technologies from the late 1970s endure today, with one notable exception: the fundamental marching orders for the vast majority of the world’s computers.
The x86 instruction set architecture (ISA), used today in more than 90 percent of the world’s PCs and servers, hit the marketplace in 1978 as part of Intel’s 8086 chip.
So when the worldwide Intel developer’s community gathers for its annual conference in Beijing later this month, they’ll spend most of their time talking about technology that was developed when Jimmy Carter was in the White House and the soundtrack for the John Travolta movie Saturday Night Fever was the best-selling album in the United States.
Other instruction sets–which are basically, lists of operations that a software program can use–do exist, of course. There’s IBM’s Power, Sun Microsystems’ Sparc and Intel’s own EPIC (explicitly parallel instruction computing) Itanium project, to name a few. But x86 continues to thrive and has no serious competitors on the horizon because it provides “good enough” performance and because of the vast amount of software written over nearly three decades. continue reading…