Intel Timna
Intel's first CPU with integrated GPU and RAM controller.
Benno Rothenberg · CC BY 4.0
The Intel Timna was a proposed central processing unit (CPU) family by Intel, announced in 1999 and designed in Haifa, Israel. It was notable as Intel's first CPU to integrate a graphics processing unit (GPU) and a random access memory (RAM) controller, intended to work with RDRAM.
- Codename
- Timna
- Announced
- 1999
- Designed in
- Haifa, Israel
- Expected clock speed
- 600 MHz to 700 MHz
- Front-side bus
- 133 MHz
- L2 cache size
- 128 KB
- Manufacturing process
- 180 nm
Lore & Background
The Timna project was announced in 1999 and designed in Haifa, Israel, named after the Timna Valley in Israel. It was intended to be Intel's first CPU with an integrated GPU and RAM controller, designed to work with RDRAM. However, the price of RDRAM did not drop as Intel expected, leading to the decision to use the Memory Translator Hub (MTH), also used by the Intel 820 chipset, to link Timna with SDRAM. A serious defect was later discovered in the MTH design, causing Intel 820-based motherboards using it to be recalled. The MTH was rebuilt but problems remained. The GPU codenamed Capitola, to be used with Timna, was based on the i752/i754 used with the 810/815E chipsets. Timna was canceled on September 29, 2000.
Reader's Guide
The Timna project's significance lies in its pioneering integration of a GPU and memory controller onto a CPU, a concept that later became standard. Its cancellation was driven by the failure of RDRAM prices to drop as expected and the defective MTH adapter, which introduced instability. The project's legacy continued through Intel Haifa, which was later tasked as the backup team for Intel's mobile CPU design. The knowledge gained from Timna's P6-derived architecture heavily influenced the Banias project, which became the foundation for Intel's successful mobile processors. Timna thus represents an early, ambitious attempt at system-on-chip integration that, despite its failure, informed future Intel designs.
Did You Know?
- The GPU for Timna, codenamed Capitola, was based on the i752/i754 used with 810/815E chipsets.
- Intel Haifa's work on Timna heavily influenced the Banias project.
From Monolithic Dies to Disaggregated Silicon
Intel's processor roadmap reveals a fundamental shift in how silicon is constructed. For years, processors were built as single monolithic dies, but this changed with Meteor Lake in late 2023, which became the first Intel mobile generation to employ a chiplet architecture, assembling the processor as a multi-chip module. Arrow Lake, released in October 2024, carried this disaggregated MCM design forward, expanding beyond mobile into desktop platforms on the LGA 1851 socket. Lunar Lake, also part of the Core Series 2 family, pushed further by integrating LPDDR5X-8533 RAM directly on-package. A notable architectural departure in this era is the elimination of Hyper-Threading in both Arrow Lake and Lunar Lake designs. Meanwhile, the value tiers within Core Series 2 actually reuse Raptor Lake-H and Raptor Lake-U silicon as re-refreshes, with Hyper-Threading disabled in the U variants, illustrating how Intel layers new architectures alongside proven older designs to fill market segments.
A New Branding Hierarchy and Enterprise Segmentation
In June 2023, Intel overhauled its processor naming, retiring the long-standing Core i3, i5, i7, and i9 labels in favor of a two-tier system: Core for everyday and value-oriented chips, and Core Ultra reserved for premium offerings. Each tier is further subdivided into 3, 5, 7, and 9 grades. This rebranding coincided with the Meteor Lake launch on December 14, 2023, and extended through subsequent generations. The product hierarchy is further stratified by power class—High Power, Low Power, and Ultra Low Power for mobile, with desktop parts on specific sockets like LGA 1700 or LGA 1851. Enterprise features such as Intel vPRO and ECC memory support are selectively applied, with vPRO Enterprise appearing on underlined premium models and ECC memory tied to specific chipsets like W680, W880, and WM790. The Stable IT Platform Program designates certain value-tier parts for long-term availability, while Intel Speed Select is offered on select Xeon 6+ models. This layered segmentation allows Intel to target everything from budget laptops to data center infrastructure under a unified but clearly differentiated brand architecture.
The 2026 Product Landscape: Data Centers, Handheld Gaming, and Refreshes
The 2026 processor landscape, as revealed through Computex and CES announcements, shows Intel pursuing distinct strategies for different markets. At Computex on June 1, 2026, the Xeon 6+ line codenamed Clearwater Forest was unveiled for network infrastructure and data center servers, supporting up to 1.5 TB of DDR5 RDIMM at 8000MT/s speeds with ECC memory, and its Accelerated Processor Server models are E-Core only. At CES on January 5, 2026, Core Series 3 arrived under the Panther Lake codename, exclusively targeting mobile processors and handheld gaming devices—a notable departure from desktop offerings. Then on March 26, 2026, Core Series 2 Plus launched as an Arrow Lake Refresh, bringing updated desktop parts to the LGA 1851 socket alongside mobile variants. The handheld gaming category within Panther Lake represents a new product segment, while the data center push with Clearwater Forest signals Intel's continued investment in server-grade silicon with massive memory bandwidth and capacity.
Stretching Proven Silicon Across Generations and Form Factors
A recurring pattern throughout Intel's modern processor history is the strategic reuse and rebranding of existing silicon across multiple product generations. Raptor Lake-H and Raptor Lake-U designs, originally from the 12th and 13th generation era, reappear as re-refreshes in the value tiers of both Core Series 1 and Core Series 2. Alder Lake parts surface in the 12th, 13th, and 14th generation lineups, sometimes as E-Core-only value options or as repurposed desktop chips for mobile use. The 14th generation, launched in October 2023 for desktop and January 2024 for mobile, is described as an iterative refresh of Raptor Lake-S and Raptor Lake-HX respectively. Even within Core Series 2, the Ultra Low Power mobile parts are based on Lunar Lake architecture while the Low Power Arrow Lake-U variants use refreshed Meteor Lake silicon on the Intel 3 node. This practice of stretching proven designs across multiple generations and form factors—desktop, mobile, embedded, and IoT—allows Intel to maintain broad market coverage while new architectures develop in parallel.
Gallery






Frequently Asked Questions
What is Intel Timna?
Intel Timna was a CPU family that Intel announced in 1999 as a proposed new line of processors. The design work was carried out at Intel's site in Haifa, Israel.
What made Intel Timna stand out from other Intel CPUs of its era?
It was the first Intel processor to place both a graphics processing unit and a memory controller directly on the same die. That integrated design was intended to pair with RDRAM for tighter memory and graphics coupling.
What were the expected specs of Intel Timna?
The chip was targeted at 600 to 700 MHz on a 133 MHz front-side bus, with 128 KB of L2 cache. Those numbers would have put it in the upper-mid consumer segment at the time.
Where was Intel Timna designed?
All of the design work was done at Intel's research and development center in Haifa, Israel. It was one of the more notable projects to come out of that particular site.
Did Intel Timna ever actually ship to customers?
No; Timna remained a proposed design and never made it to market as a commercial product. It served as an early proof-of-concept for on-die GPU and memory-controller integration that later influenced Intel's broader roadmap.
More in Intel and AMD Microprocessors, Part 3 1-24
Spotted an error? Know more?
Reader corrections go straight into our review queue. Suggest an edit · How this site is sourced
