Silicon Investor (SI) -- The First Internet Community

STOCKTALK

We've detected that you're using an ad content blocking browser plug-in or feature. Ads provide a critical source of revenue to the continued operation of Silicon Investor. We ask that you disable ad blocking while on Silicon Investor in the best interests of our community. If you are not using an ad blocker but are still receiving this message, make sure your browser's tracking protection is set to the 'standard' level.

Technology Stocks : Intel Corporation (INTC) -- Ignore unavailable to you. Want to Upgrade?

To: Scumbria who wrote (107850)	8/21/2000 10:40:48 AM
From: Gary Ng	Read Replies (1) \| Respond to of 186894

Scumbria, Re: Why a double speed ALU is almost useless And Intel just hired all those engineers to spend so much time and design something that is almost useless ? There must be something so special about working for Intel that attracts all those stupid engineers to work for it whereas all the smart guys work for some other companies :-) gary

To: Scumbria who wrote (107850)	8/21/2000 10:54:15 AM
From: Elmer	Read Replies (2) \| Respond to of 186894

Re: "Why a double speed ALU is almost useless" I am saddened to hear that if only Intel's designers had contacted you they would have known their efforts were for not. Obviously they had no idea what they were doing and just threw it in there because they thought it sounded snappy. EP

To: Scumbria who wrote (107850)	8/21/2000 10:59:04 AM
From: Tony Viola	Read Replies (1) \| Respond to of 186894

Not useless to me. I look at it this way: the ALU is used for a high percentage of CPU instruction executions. For all the fewer cycles of ALU busy time that there are because of the double speed ALU, the entire pipeline gets to take advantage of all the extra time gained. In other words, those "slower" pipeline stages still have to wait quite often for faster stages ahead of them. The faster the ALU gets those quick ALU executions the heck out of the way the faster the slower ones can advance. An analogy: toolbooths quite a few years ago put machines in that you could throw your money into, and you didn't even have to stop. These result in everyone getting through faster, the people throwing their money in, and the people that have to break a ten. Tony

To: Scumbria who wrote (107850)	8/21/2000 12:50:23 PM
From: Tenchusatsu	Read Replies (1) \| Respond to of 186894

Scumbria, <If they made the speed limit on one mile of the Bay Bridge 100MPH, would that allow more people to get through the toll booth each hour?> You just described Flynn's Bottleneck (named after a guy who made money off of stating the obvious). The number of instructions you execute is directly limited by the number of instructions you fetch. No self-respecting CPU architect would ever forget about that bottleneck. In my opinion, there are two uses for a double-speed ALU: 1) You only need to implement half as many execution pipes. 2) Certain data hazards can be bypassed. If there is an instruction that depends on the results of the previous instruction, both can be sent down the double-speed ALU and be executed in one (regular) clock cycle. Two execution units running at normal speed will need two clock cycles to handle those instructions (thereby leaving one unit unutilized). Tenchusatsu