Q2 2007
AMD Opteron 2222 SE was released on April 5th. The 822SE is a dual core processor featuring a clock speed of 3.0GHz.

AMD Opteron 8222 SE was released on April 5th. The 822SE is a dual core processor featuring a clock speed of 3.0GHz.

AMD CPU Price cuts occurred on April 9th. See the AMD CPU Prices page for more information

Intel Core 2 Quad QX6800 was released on April 9th. The QX6800 features 2 x E6800 cores in the same package, comprising a quad core CPU running at a clock speed of 2.93Ghz, 2x4MB L2 cache and running on a 1066Mhz FSB. The QX6800 will initially retail for $1199.

nVidia GeForce 8600 (G84) GPU was released on April 17th. G84 is a mid range GPU and the successor to the GeForce 7600, featuring the enhancements found in the G80 core but with 32 stream processors and a 128-bit memory interface. In addition, G84 features a new video processor with dedicated circuitry for HD video decoding and a BSP engine. The 8600 series features the following boards:

GeForce 8600GTS - 675Mhz core, 1.45Ghz stream processors, 2.0Ghz memory, 71W - $199-229
GeForce 8600GT - 540Mhz core, 1.15-1.18Ghz stream processors, 1.4Ghz memory, 43W - $149-159

Initial benchmarks show that the 8600GTX performs very slightly better than a GeForce 7900GT, and therefore slightly slower than the similarly priced Radeon X1950 Pro, in current DirectX 9 games. However, 3DMark06 performance is good, suggesting that future games utilising more shader intensive code, may perform better than the current mid range DX9 cards.

nVidia GeForce 8500 (G86) GPU was released on April 17th. The GeForce 8500 is based on the G86 core - a cut down version of the G84 core - featuring 16 stream processors and a 128-bit memory bus. Like the G84, the G86 features a new video processor with dedicated circuitry for HD video decoding and a BSP engine. The G86 series features the following boards:

GeForce 8500GT - 450Mhz core, 900Mhz stream processors, 1.4Ghz memory - $79-99
GeForce 8400GS - 450Mhz core, 900Mhz stream processors, 800Mhz memory (OEM)
GeForce 8300GS - 450Mhz core, 900Mhz stream processors, 800Mhz memory (OEM)

Intel CPU Price cuts occurred on April 22nd. See the Intel CPU Prices page for more information.

Intel Core 2 Duo E6420 was released on April 22nd. The E6420 is expected to feature a clock speed of 2.13GHz running on a 1066Mhz FSB and featuring 4MB (compared to 2MB in the E6400) of L2 cache. The extra L2 cache provides roughly a 0-10% performance improvement (with an average of around 2-3%), dependant upon the application. The E6420 is expected to retail for $183.

Intel Core 2 Duo E6320 was released on April 22nd. The E6320 is expected to feature a clock speed of 1.83GHz running on a 1066Mhz FSB and featuring 4MB (compared to 2MB in the E6300) of L2 cache. The extra L2 cache provides roughly a 0-10% performance improvement (with an average of around 2-3%). The E6320 is expected to retail for $163.

Intel Core 2 Duo E4400 was released on April 22nd. The E4400 is expected to feature a clock speed of 2.0GHz running on an 800Mhz FSB and featuring 2MB of L2 cache. The E4400 is expected to retail for $133.

nVidia GeForce 8300 (G82) GPU is expected to be released at the End of April. G82 is a low end GPU and the successor to the GeForce 7200, featuring the enhancements found in the G80 core. The GeForce 8600 is expected to be notably faster than it's predecessor the GeForce 7300 as well as being more feature rich.

GeForce 8300 GT - 32 shader processors, 500MHz core, 1.2GHz memory, 128/256MB - $89-99
GeForce 8300 GS - 24 shader processors, 500MHz core, 1GHz memory, 128/256MB - $69-79


ATI Radeon HD 2900 (R600) GPU is expected to be released in May on an 80nm process. The R600 is the successor to the R580 core, featuring 700M transistors. R600 will be fully DirectX 10 compliant, utilising a Unified Shader Model architecture. Current rumours suggest that R600 will feature 320 stream processors with 32 TMUs and 32 ROPs running at a clock speed of around 800Mhz (possibly 814Mhz) for the XTX version. The R600 core is also expected to feature support for 128-bit HDR, physics processing, 24x AA and integrated HDMI outputs with 5.1 surround sound support. R600 will also feature ATI's Avivo HD technology, which is provided by the new Universal Video Decoder (UVD), providing hardware acceleration for H.264 and VC-1 decoding, as used by Blu-ray and HD DVD titles. R600 is expected to interface to GDDR3 or GDDR4 Memory over a 512-bit interface. R600 is reported to consume up to 270W (twice that of R580) and will require two PCI Express power connectors. R600 is expected to be slightly faster than nVidia's GeForce 8800 GTX and will be an exceptionally large graphics card with a very large cooling system. There are likely to be three parts released:

Radeon HD 2900 XTX - 12" / 9.5" board, 1GB GDDR4, Dual Slot (12" OEM - 270W, 9.5" Retail - 240W)
Radeon HD 2900 XT - 9.5" board, 750Mhz core, 1.64Ghz memory, 1GB or 512MB GDDR4, Dual Slot (240W). $399
Radeon HD 2900 XL - 9.5" board, 512MB GDDR3, Single Slot (Released at a later date)

Pre-release benchmarks suggest that the Radeon HD 2900XT will around 10-20% faster than nVidia's GeForce 8800GTX in 3DMark 2006.

ATI Radeon HD 2600 (RV630) is expected to be released in May on a 65nm process. The RV630 is the mid-range version of the R600 GPU interfacing to 256MB or 512MB of 128bit GDDR4 memory. The RV630 also features integrated HDMI & HDCP support and offer integrated audio functionality. The GPU also features support for Crossfire and PCIe 2.0 for increased interface bandwidth. The initial members of the family are expected to be as follows:

Radeon HD 2600 XT - ? GPU clock, ?MHz GDDR4 memory, 256/512MB, 121/128W TDP
Radeon HD 2600 Pro - ? GPU clock, ?MHz GDDR3 memory, 256MB, <93W TDP
Radeon HD 2600 - ? GPU clock, ?MHz DDR2 memory, 256/512MB, <75W TDP

ATI Radeon HD 2400 (RV610) is expected to be released in May on a 65nm process. The RV610 is the entry-level version of the R600 GPU interfacing to 128MB or 256MB of 64bit DDR2 memory. Like the RV630, RV610 feature support for HDMI & HDCP with integrated audio functionality. The initial members of the family are expected to be as follows:

Radeon HD 2400 XT - ? GPU clock, 700MHz GDDR3 memory clock, 128/256MB, 35W TDP
Radeon HD 2400 Pro - ? GPU clock, 400MHz DDR2 memory clock, 64-bit 128/256MB, <25W TDP

GeForce 8800 Ultra is expected to be released in May to compete with ATI's R600 GPU. The GeForce 8800 Ultra is based around the G80 core and is expected to be a higher clocked version of the 8800GTX, offering an expected 10-15% performance improvement (i.e. core clocks of around 650-700Mhz).

nVidia GeForce Go 8800GS (G80M) is expected to be released in May. The GeForce Go G80M core is the mobile equivalent of the G80 GPU, with the GS model being a lower clocked version of the GTX due subsequent release.


Intel Core 2 Quad Q6400 is expected to be released in June. The Q6400 is expected to feature a clock speed of 2.13GHz running on a 1066Mhz FSB and featuring 2x4MB of L2 cache.

Intel Pentium E2000 series budget CPUs are expected to be released in June. The E2000 series is based around the Dual Core Conroe design, but with 1MB of L2 cache disabled and Intel's Virtualization Technology (VT), ViiV and vPRO functionality disabled. The first members of the E2000 family are expected to be:

Pentium E2160 - 1.8 GHz, 800Mhz FSB, 1 MB of L2 cache, $84
Pentium E2140 - 1.6 GHz, 800Mhz FSB, 1 MB of L2 cache, $74

Intel Celeron 400 series budget CPUs are expected to be released in June. The Celeron 400 series are single core version of the Pentium E2000 series, featuring an 800Mhz FSB and 512KB L2 cache. The first members of the E2000 family are expected to be:

Celeron 440 - 2.0Ghz, 800MHz FSB, 512KB L2 cache, $59
Celeron 430 - 1.8Ghz, 800MHz FSB, 512KB L2 cache, $49
Celeron 420 - 1.6Ghz, 800MHz FSB, 512KB L2 cache, $39

nVidia GeForce Go 8800GTX is expected to be released in June. The GeForce Go features the G80M core first released in the 8800GS, with the GTX being a higher performance part.

Q2

AMD Athlon X4? (Barceona) is expected to be released in Q2. Barcelona is AMD's first quad core CPU based on a the 65nm process. Barcelona additionally introduces the Revision H core for quad processors.

nVidia GeForce Go G83M is expected to be released in Q2. The GeForce Go G83M is the mobile equivalent of the G83 GPU.

nVidia GeForce Go G82M is expected to be released in Q2. The GeForce Go G82M is the mobile equivalent of the G82 GPU.

ATI Radeon Mobility M78 is expected to be released in Q2. This GPU is based on the R600 core and is targeted at the high end of the mobile market.

ATI Radeon Mobility M76 is expected to be released in Q2. This GPU is based on the R600 core and is targeted at the mid range mobile market.

ATI Radeon Mobility M72 is expected to be released in Q2. This GPU is based on the R600 core and is targeted at the low end mobile market.

H1

AMD Opteron processor, based on the Opteron+ core, is expected to be released in Q2 at a clock speed of 2.8Ghz.

AMD Athlon X2 (Brisbane) is expected to be released in H1. Brisbane is the 65nm successor to Windsor and introduces the K8 Revision G core. Brisbane features either 2x1MB L2 cache or 2x512KB L2 cache and the Brisbane core will effectively replace all Athlon X2 and Opteron Dual Core Revision E and Revision F processors with a single architecture.

AMD Turion 64 (Tyler) is expected to be released in H1. Tyler is the successor to Taylor and will be built on a 65nm process. Tyler is a Dual Core Socket S1 Mobile CPU featuring 2 x 512KB L2 cache, DDR2 800 support and a 35W TDP.

AMD Mobile Sempron (Sherman) is expected to be released in H1. Sherman is the successor to Keene and will be built on a 65nm process. Sherman is a Dual Core Socket S1 Mobile CPU featuring 2 x 512KB L2 cache, DDR2 800 support and a 25W TDP.

Q3 2007
AMD Opteron (Barcelona) Quad Core CPU is expected to be released in July on a 65nm process. Barcelona is expected to be the first K8L based CPU, featuring 4 cores. Each core will feature 64KB - 32KB Instruction, 32KB Data - of L1 (down from 128KB in the K8 architecture), 512KB of L2 cache per core and - in it's Opteron form, 2MB of shared L3 cache. K8L will also feature AMD's DICE (Dynamic Independent Core Engagement) power saving technology which enables each core to alter it's own p-state (Power state) right down to putting a core in a full Halt condition and will introduce HyperTransport 3. Hypertransport 3 will introduce a number of improvements. Firstly, the HT speed will be increased to 2.6Ghz, which will allow for 5.2GT/s, compared with a maximum of 1.4Ghz in HT2 (1Ghz in the K8 architecture). Secondly HT3 will introduce 'Un Ganging', which will allow either one 16-bit link or two 8-bit links to be created on the fly. This will be particularly useful with multi-socket Opteron servers as it can allow for single memory hop access to memory which would previously have taken two hops. Additionally the K8L core will have an enhanced instruction set, Indirect branch prediction, 32-byte prefetch (compared with 16 in the K8 architecture), 48-bit addressing with 1GB pages, better cache coherency, I/O virtualisation, Memory mirroring, data poisoning and HT retry protocol support, and 2x128-bit SSE units (compared with 2x64-bit units in K8) featuring support for single cycle 128-bit instructions. Altair will interface to DDR2 memory, with the K8L core featuring support for FBD and, in a future memory controller revision, DDR3 and FBD2 support. The initial members of the family will be:

Opteron 2258HE - Quad Core, 68W
Opteron 1266 - Dual Core, 95W
Opteron 1268SE - Dual Core, 120W
Opteron 1270SE - Dual Core, 120W

nVidia GeForce 8900 (G81) is expected to be release in June. The featureset of this part is currently not known, although it is expected to feature higher clock speeds than it's predecessor the 8800 series.


AMD Athlon 64 FX (Agena), previously known as Altair, is expected to be released in Q3 on a 65nm process. Agena is the desktop equivalent of Barcelona, being a Quad Core K8L CPU. Agena will be a HyperTransport 3.0 CPU running on Socket F+.

AMD Athlon 64 X2 (Kuma), previously known as Antares, is expected to be released in Late Q3. Kuma is the Dual Core, Socket AM2+ version of Agena, featuring HT3 support and 1MB (possibly 2MB) of L3 cache.

nVidia MCP72 single-processor chipset for the AMD HT-3 / AM2+ architecture is expected to be released in Q3 alongside Agena. MCP72 will feature support for HyperTransport 3.0 and PCIe 2.0, and will additionally feature six SATA 3.0Gbps and dual Gigabit Ethernet ports.

Intel Stealey mobile CPU is expected to be released in Mid 2007. Stealey is expected to be a single core budget CPU featuring 512KB of L2 cache.

Intel Core 2 Quad (Yorkfield) CPU is expected to be released in Q3 and is expected to be based on a 45nm process. Yorkfield will be the successor to Kentsfield and feature two 6MB L2 caches (one cache for two cores), making a total of 12MB L2 (2x6MB). Yorkfield is also expected to feature the 50 additional Penryn New Instructions (SSE4) and feature a clock speed of 3.46 to 3.73Ghz. Yorkfield will be paired with the Bearlake chipset family and will feature a 1333Mhz or 1066Mhz FSB speed.

Intel Core 2 Duo (Wolfdale) desktop CPU is expected to be released in Q3. Wolfdale is the Dual Core version of Yorkfield based on the Penryn core and featuring 6MB of shared L2 cache running on a 45nm process. Wolfdale is expected to feature clock speeds of up to 4Ghz and can run on a 1333Mhz or 1066Mhz FSB speed. The TDP of Wolfdale is expected to be 57W.

Intel P35 Express (Bearlake-P) chipset for the Core architecture is expected to be released in Q3. Bearlake-X is the successor to Intel's 965P chipset. Bearlake P is expected to feature support for DDR3 1066 and DDR2 800 memory as well as a 1333Mhz FSB speed. Bearlake P will also feature Intel's ICH9 southbridge.

Intel G35 Express (Bearlake-G) chipset for the Core architecture is expected to be released in Q3. G35 is the integrated graphics version of P35, featuring an integrated DirectX 9.0c GPU core.

Intel X38 (Bearlake-X) chipset for the Core architecture is expected to be released on Q3. Bearlake-X is the successor to Intel's 975X chipset at the high end of the market and is expected to feature PCI Express 2.0 (running at a 5Ghz rather than 2.5Ghz clock), 2xPCIe x16 ports, support for a 1333Mhz PSB and PC2-6400 (DDR2 800) and PC3-1333 (DDR3 1333) memory support. Bearlake X will also feature Intel's ICH9 southbridge.

Intel Bearlake G+ chipset for the Core architecture is expected to be released on Q3. Bearlake-G+ is the integrated graphics version of Bearlake X, featuring an integrated DirectX 10 GPU with HDCP support.

ATI RD790 chipset is expected to be released in Q3 alongside Altair and Antares. RD790 is expected to feature 2x16 or 4x8 PCI Express ports (PCI Express Generation 2) and support for Hypertransport 3. RD570 will be released alongside the SB700 south bridge, featuring 6xSATA300 ports, 12xUSB2 ports and 'flash support'.

ATI RX790 chipset is expected to be released in Q3. RX790 is a cut down version of RD570, featuring support for a single graphics card.

ATI RS790 chipset is expected to be released in Q3. RS790 is the Integrated Graphics version of RX790, featuring full DX10 support.

ATI RS740 chipset is expected to be released in Q3. RS740 is the DirectX 9 Integrated Graphics version of RS790.

Serial ATA 3 is expected to be introduced in Mid 2007, offering a data transfer twice as high as SATA 2, i.e. 600MB/s.

WinFS RTM is expected to occur in Q3.

Intel Xeon DP 2.6Ghz (Woodcrest) processor is expected to be released in Q2.

Windows Server Longhorn RTM is expected to occur in Q3.

Intel 45nm process is expected to come online in Q3.

Intel Penryn architecture CPUs are expected to be released in Late Q3 on a 45nm process. Penryn is an update to the Core 2 architecture, designed for both mobile and desktop CPUs. Enhancements to the Penryn architecture include larger L2 caches (6MB for dual-core CPUs and 12MB for quad-core), faster division operations (twice the speed), improved cache management, enhanced power management including a C6 'Deep Power Down' state, support for up to a 1600Mhz PSB, Enhanced Dynamic Acceleration technology, and SSE4 support with an enhanced SSE engine. Intel's Enhanced Dynamic Acceleration technology allows heavily utilised cores to increase their clock speed whilst staying within a specified thermal and power envelope. SSE4 introduces 50 new instructions to the core for streaming operations, enhancing the performance of media and high performance computing applications.

Intel's performane metrics suggest roughly a 20% increase in gaming performance (3.2Ghz Penryn vs 2.93Ghz Core 2) and up to 40% with codecs that utilise SSE4. Virtual machine transition performance is also improved by 25-75%.

Intel Seaberg chipset for Xeon Penryn CPUs is expected to be released in Late Q3. Seaberg is expected to feature a 1600Mhz PSB, up to 128GB of FB-DIM memory, a 16MB+ Snoop filter and 10Gb Ethernet.

Q4 2007
AMD Sempron (Sparta) is expected to be released in H2 2007. Sparta is the 65nm successor to Manila and will be based on the AM2 platform.

AMD Opteron (Deerhound) is expected to be released in H2 2007 on a 65nm process and running on a Socket F platform. Deerhound will be a 4 core CPU featuring a shared L2 cache and interfacing to Dual Channel Registered DDR2 memory.

AMD Athlon 64 X2 (Budapest) is expected to be released in Q4. Budapest is expected to be the HT2 version of Kuma for Socket AM2.

AMD Athlon 64 X2 (Rana), formally known as Arcturus, is expected to be released in Q4. Arcturus is a dual core CPU based around the Kuma core, but featuring no L3.

AMD Spica is expected to be released around Q4 07 / Q1 08. Spica is the single core version of the K8L core.

Windows Server Longhorn is expected to be released in Q4.

Windows Vista SP1 is expected to be released in Q4 alongside Windows Server Longhorn. SP1 is expected to feature a number of enhancements as well as security updates, and is expected to update the kernel to bring it up to date with that found in Windows Server Longhorn. Enhancements include the support for UEFI (Unified Extensible Firmware Interface), which will replace the current BIOS architecture of new PCs, and an updated version of Media Center.

Microsoft Visual Studio 'Orcas' is expected to be released in Q3/Q4. Visual Studio 'Orcas' is the successor to Visual Studio 'Whidbey'. Visual Studio 9 will provide support for Vista's enhanced UI, managed interfaces, new trustworthy computing and security models, new application model, improved communication and collaboration, integrated data storage, and innovations in presentation and media.

Intel Tolapai system on a chip CPU is expected to be released in Q4 on a 65nm process. Tolapai will feature a Pentium-M derived CPU core alongside a north and southbridge (but no GPU) on a single chip. Tolapai will also feature hardware accelerated security encryption and decryption functions, similar to those found on VIA's C7 and C3 Nehemiah processors, including AES, 3DES, RC4, MD5, SHA-1, SHA-224, SHA-256, SHA-384, SHA-512, HMAC, ESA and DSA. The core will feature 256KB L2 and will run at 600Mhz, 1066Mhz or 1200Mhz with a 13-22W TDP. The chip will support up to 2GB DDR2 400-800 and up to 3xGb ethernet ports (with a PHY). Also supported by the built-in southbridge is FDD support, Parallel & Serial ports, PS2 ports and 2xSATA and 2xUSB2 ports.

2007

Intel Millville desktop CPU is expected to be released in Q1 2007. Millville is expected to be a single core version of Allendale, containing 1MB of L2 cache.

Intel Xeon MP (Tigerton) CPU is expected to be released in 2007 as part of the Caneland platform. Tigerton is the replacement of the now abandoned Whitefield CPU and is expected to be a 4 core CPU possibly based on the Merom core and using Intel's CSI bus - akin to AMD's HyperTransport. Like Whitefield, Tigerton is being designed in Intel's design centre in India.

Intel Tukwila (formally known as Tanglewood, possibly cancelled) IA64 CPU is expected to be introduced in 2007. Tukwilla is the multi-core (4 core) successor to Shavano (now probably successor to Montecito) built on a 65nm process. Tukwila is expected to contain 32Mb L3 cache.

Intel Dimona IA-64 CPU is the Dual Processor only version of Tukwilla, expected to be released in 2007.

Intel Gilo Mobile processor, successor to Merom, is expected to be released in 2006/7 built on a 65nm process. Gilo is expected to be a multi-core CPU.

Sun Rock CPU is expected to be released in 2007 on a 65nm process. The successor to Niagara, this processor is expected to be extreme CMT design with performance expected to be up to 30x more throughput than a 1Ghz USIII processor.

Sun Niagara 2 CPU, the successor to the UltraSPARC T1 (Niagara), is expected to be released in 2007 on a 65nm process. It will be the cheaper version of Rock and an evolution of the Niagara processor. Niagara 2 is expected to feature 8 cores, as did Niagara, but the architecture is significantly redesigned. Each core will be able to execute up to 8 threads (compared to 4 in Niagara) with up to 2 instructions per cycle (compared to 1 instruction per cycle in Niagara) and there will be one FPU per core rather than having one FPU for all 8 cores as seen in Niagara. The FPU capabilities of Niagara 2 will additionally be helped by significantly reduced latencies, with raw floating performance increased by up to 50X in practice. The CPU will also feature enhanced support for cryptographic functions, with Niagara 2 being able to assist with many more types of cryptography, such as elliptic curve cryptography, in addition to the DSA and RSA support featured in Niagara 1. The CPU will also feature on-die PCI-Express connections for general I/O, multiple on-die 10Gbit ethernet connections with TCP/IP offloading and a simple SMP interface, allowing for low-end SMP scalability. The CPU will interface to FB-DRAM via a 50GB/s+ memory interface.
Niagara 2 is expected to run at a clock speed of around 1.4GHz, compared to 1.2GHz for the UltraSPARC T1, with overall performance expected to be at least double that of Niagara 1.