Specs for AMD’s Bulldozer CPUs Revealed

It seems like AMD is experiencing some difficulty with the yields of their chips that can hit clockspeeds of revisions B0 and B1, but we’re not certain whether it’s the base clockspeed or Turbo Core based on sources. Here’s the list of processors that the company is launching:

  • FX-8150: 3.6GHz (4.2GHz Turbo Core), 8-core, 8MB L2 cache, 125W
  • FX-8120: 3.1GHz (4GHz Turbo Core), 8-core, 8MB L2 cache, 125W/95W
  • FX-8100: 2.8GHz (3.7GHz Turbo Core), 8-core, 8MB L2 cache, 95W
  • FX-6120: Clockspeeds TBD, 6-core, 6MB L2 cache, 95W
  • FX-6100: 3.3GHz (3.9GHz Turbo Core), 6-core, 6MB L2 cache, 95W
  • FX-4120: Clockspeeds TBD, 4-core, 6MB L2 cache, 95W
  • FX-4100: 3.6GHz (3.8GHz Turbo Core), 4-core, 4MB L2 cache, 95W

All of these CPUs will have an 8MB of L3 cache and take DDR3-1866 memory, and something to note is that they’re based on a 32nm process. We should see them surface around August or September with varying time frames for the parts.


AMD Phasing Out Phenom II X6 Series

According to news sources, the chip maker has plans to mark their entire Phenom II X6 processor line to EOL (end of life) by the fourth quarter of this year. This is a move to put the spotlight on the new FX series of CPUs based on Bulldozer. The Phenom II X6 1045T, 1055T and 1065T are the first chips marked for retirement, but can still be ordered until third quarter (2011).

Even though AMD won’t manufacturing these chips, however they will still honor any warranty claims that are still valid.

Early Benchmarks for AMD’s “Llano” Platform Surface

Apparently somebody got their hands on the AMD A8-3800 Quad-Core APU along with Gigabyte’s GA-A75-UD4H motherboard and they did exactly the right thing. Ran several benchmarks. Aside from the two feature parts, other parts used were 4GB (2x2GB) of GSKILL DDR3-1600 RAM, a standard 1TB Seagate Barracuda 7200.12 HDD, and note that the graphics is the on-board Radeon HD 6550D that’s on the same die.

What we found fairly darn impressive (for an integrated solution) were the benchmarks in the gaming department with these scores (at 1080p):

  • Street Fighter 4: 50.32 FPS
  • Hawx: 54 FPS (DX9) / 22 FPS (DX10)
  • Hawx 2: 46 FPS (DX9) / 34 FPS (DX11)
  • Resident Evil 5: 29.0 FPS (DX9) / 27.4 FPS (DX10)

Brand New Pentium Chips, Heading to a LGA 1155 Socket Near You

Ah, the Pentium brand that we all came to know since two decades or so and it looks like Intel isn’t retiring it any time soon. In fact they’re getting ready to launch two new chips based on the Sandy Bridge architecture within this month.

The three slated chips will support dual-channel memory and are fully compatible with LGA 1155 boards. At the lower end of the spectrum is the Pentium G620, as it’s only a dual-core CPU clocked in at 2.6GHz and contains 3MB of cache – will retail for about $65. Following is the more powerful G840, and it’s also a dual-core but clocked at 2.8GHz with the same 3MB cache – priced at $75. Finally the best of the bunch is the G850 having the highest clock of them all at 2.9GHz, but sadly none of these new parts feature Turbo Boost.

Although official launch is Q2 of this year, but keep an eye out as these new chips may surface and appear on retail shelves as soon as this week.

All the Answers to Intel’s Brand New Z68 Chipset – Part 2 (Feature)

Looking back at our last article, we’ll continue with discussing Intel’s SRT (Smart Response Technology) and this cool feature on the Z68 chipset uses an SSD to cache more frequently used data from an hard drive. One can expect up to a 4x increase in performance over the traditional HDD.

The process of setting up this new tech requires a few simple steps to get it up and running, as you need to ensure that both the SSD and hard drive are hooked up then you need to make sure that the Intel controller default of AHCI is set to RAID in UEFI. After that you can go about installing Windows 7 to your hard drive as normal, but you can’t enable SRT until all the drivers are installed. Finally it’s just a mater of hitting the Accelerate button in the RST driver (see below).

Onwards, simply select the drive you’d like to use as a cache and select the disk you want “accelerated” (would usually be C:\) then choose the mode of Enhanced or Maximized (see following image). First option Enhanced is probably your best bet in terms of being safer, and it offers better read speeds but write speeds remain the same as the HDD used. Second is Maximized and this is a form of caching since data is written to the SSD then synced to the hard drive. This mode should yield similar write speeds as what the SSD used is capable of.

At the end of the day, SRT is ideal for those who can’t afford large capacity SSDs but want that solid state performance. By leveraging this tech, individuals can get smaller SSDs and be able to achieve similar speeds when used in conjunction with a traditional hard drive.

All the Answers to Intel’s Brand New Z68 Chipset – Part 1 (Feature)

At the a first glance there aren’t a whole lot of changes opposed to its predecessor P67, but there are some significant improvements that make the Z68 platform worth considering. Let’s jump straight in, and they include:

  • Being able to overclock the graphics core in the CPU (Sandy Bridge)
  • A new IPT (Identity Protection Technology) feature that integrates a hardware token in to the PC
  • Support for switchable graphics between external graphics and the integrated solution in Sandy Bridge
  • SSD caching that uses a smaller capacity SSD and traditional hard drive to increase responsiveness

We have to agree with many critics out there that the first two won’t be the main selling points for a typical consumer, but being able to switch between might be more of an interest. The early adopters of the P67 chipset couldn’t utilize the on-die graphics simply because the board didn’t have any outputs as those were only on the H and Q-series chipsets. For the Z68 now, Intel bundles the Flexible Display Interface which allows processor graphics to the chipset and a good example which includes bother DVI and HDMI ports would be ASUS’s P8Z68-V Pro.

Here’s the part that will intrigue most people, since motherboard makers will be bundling LucidLogix’s Virtu tech as we previously mentioned. Two modes available are i-mode and d-mode, we’ll elaborate further.

When in i-mode, you are mainly utilizing integrated graphics and only tapping in to your discrete card for gaming and there is some prep that needs to be done. This includes initializing the integrated graphics port in UEFI, connecting your monitor to the motherboard’s graphics port, and installing graphics for both the on-board Intel graphics portion then your choice of the discrete card. It’s not all that labor intensive with just a couple of straight forward steps, but the problem arises when you need Lucid to make profiles for any game you decide to run in the Virtu mode and we know that several gamers are eager to dive in to the newly released title. Asides from that, power savings are insignificant, and in contrast to a mobile version of switchable graphics the external video card doesn’t shut off completely. Despite all the power management features in high-end graphics solutions today, they still consume quite a bit of power even when it’s idling. The i-mode is also not compatible at the moment with dual-GPU cards and SLI.

As for d-mode? We think that this will be the more useful one of the two. You are start by setting UEFI to initialize the PCI-E graphics adapter first then hooking up your monitor to the graphics card. In this particular mode, the external graphics solution is the main one and games can run without the use of any profiles coming from Lucid. Then you may be wondering now about the usefulness of this mode, well it’s being able to take advantage of the Quick Sync technology built right into Sandy Bridge. It may not shine when it comes to gaming, but it means serious business for encoding and transcoding since Intel dedicated transistors just for those jobs. How good is this implementation you would ask?

In d-mode, Cyberlink MediaEspresso 6.5 took 142 seconds on a GeForce GTX 580 to transcode a single VOB file to a generic WMV file. While the QuickSync mode on the Core i7-2600K took only 109 seconds, and that’s about a 30% difference. If something were to take hours, which option would you rather have?

Another rad feature is Intel’s SRT (Smart Response Technology) which lets the Z68 chipset to use an SSD to cache commonly used data from a hard drive. This supposedly yields up to a 4x improvement in performance over a traditional drive alone. Unfortunately our coverage got longer than we expected, and we’ll discuss SRT in the next segment as our two-part series. Stay tuned.

SandForce Intros 2nd Gen SSD CPUs

Looks like they officially came around announcing the new SF-2200/2100 SSD processors, after both recent product releases coming from Patriot and OCZ. Here’s what they have to say:

“With high profile products now incorporating SSDs as standard storage media and most other system vendors offering them as options, the market for client SSD applications is poised for growth as SSD prices decline,” said Joseph Unsworth, Research Director, NAND Flash & SSD at Gartner. “SSD controllers that can deliver superior performance and reliability without the dependence on DRAM will have a compelling value proposition across a wide range of client applications.”

If you guys aren’t aware, the SF-2200 yields read/write speeds of up to 500MB/s while it’s sibling caps out at 250MB/s. Both logical processors, however, comes in a single-chip “DRAM-less” factor and has DuraClass Technology (RAISE, DuraWrite), have support for complex 30nm and 20nm class NAND flash from all major vendors, advanced ECC engine correcting, power and performance optimization and tuning features, OPAL security with 256-bit AES encryption and automatic, line-rate double encryption with drive-level password.

ETA? Well, at the time of writing this – only Patriot and OCZ announced products so far with these SSD processors. More specifically the OCZ Vertex 3 and Vertex 3 Pro is to ship after CeBIT this week, while Patriot’s Wildfire series should arrive sometime in Q2 this year.

Intel Gives Preview of 32nm Itanium “Poulson” CPU

Although this isn’t exactly something compared to the “decacore” (10-core) processor they’re launching, last week at the ISSCC (International Solid-State Circuits Conference) – Intel revealed their next-gen Itanium processor dubbed “Poulson”. The new octal-core CPU is one of the most notable updates to the Itanium brand Intel has ever made and this could cause ripples in terms of performance in the high-end server/mainframe spectrum of the market. Not too shabby we’d say, since this line wasn’t well received when it first launched a decade ago and as hardware enthusiasts, we can only hope that these high end parts will eventually trickle down to the consumer level.

Intel Shipping “Decacore” Processors by Q3

We’re clueless whether this is the term they’ll use for 10-core CPUs, but Intel’s Westmere-EX Xeon processor will be shipping this half of 2011. Last week an Intel rep said these chips will follow their 2010 release of the Nehalem-EX, which is currently their fastest server processor offering on the market that has eight cores. The new Xeons also introduce brand new security features including AES-NI (Advanced Encryption Standard-New Instructions) for faster encryption and decryption processes. The CPU is based on the 32nm manufacturing process opposed the 45nm process that the Nehalem-EX used.

The Intel spokesperson also stated that the new Xeon processors are intended for high-end servers (up to eight sockets) in data centres which manage large databases and run other intensive applications. Each core will be able to run two threads, and yield a total of up to 160 threads simultaneously (8 CPUS, 20 threads each). Two-socket systems with the Westmere-EX chips can also handle up to 2TB of memory. Good news for customers is that the new part is fully backwards compatible, and will work with current Nehalem-EX configurations.

AMD Refreshes Pricing for Phenom II CPUs

Maybe it’s to ensure a strong first quarter, or maybe they’re leveraging Intel’s mishap to their advantage. Who knows? Anyways, AMD went and gone to slashing the prices of many processors in their Phenom family (select chips) by up to $30. Starting with AMD’s flagship CPU, the Phenom II X6 1100T, was dropped from $265 to $239. We’re also seeing the 1090T dip from $235 to $205, and the 1055T drop down from $199 to $175. The list goes on with the inclusion of the ever so popular quad-core processors, see the table below.

Processor Old price New price
Phenom II X4 840 $105
Phenom II X4 955 BE $145 $135
Phenom II X4 965 BE $165 $155
Phenom II X4 970 BE $185 $175
Phenom II X4 975 BE $195
Phenom II X6 1055T $199 $175
Phenom II X6 1065T $185
Phenom II X6 1075T $199 $195
Phenom II X6 1090T BE $235 $205
Phenom II X6 1100T BE $265 $239

Source: The Tech Report