摘要:
A dynamic memory requires refreshing to retain data in its memory cells. This may cause access to the dynamic memory for purposes other than refreshing (external access) and access to it for refreshing to compete with each other, resulting in a performance deterioration. According to this invention, a pipelined dynamic memory (PDRAM) is used, and the pipeline frequency (CLK) of the pipelined dynamic memory is made higher than the frequency (CLK1) of external access, and access required for refreshing is made to an unoccupied slot (a timing when any external access request is never issued) in the pipeline of the pipelined dynamic memory. This makes refreshing of the internal dynamic memory an internal operation, which eliminates the need to take refreshing into consideration at the time external access is made, leading to improvement in operating ease and speed.
摘要:
To enhance the speed of first access (read access different in word line from the previous access) to a multi-bank memory, multi-bank memory macro structures are used. Data are held in a sense amplifier for every memory bank. When access is hit to the held data, data latched by the sense amplifier are output to thereby enhance the speed of first access to the memory macro structures. Namely, each memory bank is made to function as a sense amplifier cache. To enhance the hit ratio of such a sense amplifier cache more greatly, an access controller self-prefetches the next address (an address to which a predetermined offset has been added) after access to a memory macro structure so that data in the self-prefetched address are preread by a sense amplifier in another memory bank.
摘要:
To enhance the speed of first access (read access different in word line from the previous access) to a multi-bank memory, multi-bank memory macro structures are used. Data are held in a sense amplifier for every memory bank. When access is hit to the held data, data latched by the sense amplifier are output to thereby enhance the speed of first access to the memory macro structures. Namely, each memory bank is made to function as a sense amplifier cache. To enhance the hit ratio of such a sense amplifier cache more greatly, an access controller self-prefetches the next address (an address to which a predetermined offset has been added) after access to a memory macro structure so that data in the self-prefetched address are preread by a sense amplifier in another memory bank.
摘要:
In an information processing apparatus involving a cache accessed by INDEX and TAG addresses, accesses to the main memory include many accesses attributable to the local character of referencing and write-back accesses attributable to the replacement of cache contents. Accordingly, high speed accessing requires efficient assignment of the two kinds of accesses to banks of the DRAM. In assigning request addresses from the CPU to different banks of the DRAM, bank addresses of the DRAM and generated by operation of the INDEX field and the TAG field so that local accesses whose INDEX varies and accesses at the time of writing back of which INDEX remains the same but TAG differs can be assigned to different banks. High speed accessing is made possible because accesses to the main memory can be assigned to separate banks. Furthermore, as reading and writing at the time of writing back can be assigned to a separate bank, pseudo dual-port accessing is made possible with only one port, resulting in higher speed write-back accessing.
摘要:
In an information processing apparatus involving a cache accessed by INDEX and TAG addresses, accesses to the main memory include many accesses attributable to the local character of referencing and write-back accesses attributable to the replacement of cache contents. Accordingly, high speed accessing requires efficient assignment of the two kinds of accesses to banks of the DRAM. In assigning request addresses from the CPU to different banks of the DRAM, bank addresses of the DRAM and generated by operation of the INDEX field and the TAG field so that local accesses whose INDEX varies and accesses at the time of writing back of which INDEX remains the same but TAG differs can be assigned to different banks. High speed accessing is made possible because accesses to the main memory can be assigned to separate banks. Furthermore, as reading and writing at the time of writing back can be assigned to a separate bank, pseudo dual-port accessing is made possible with only one port, resulting in higher speed write-back accessing.
摘要:
To enhance the speed of first access (read access different in word line from the previous access) to a multi-bank memory, multi-bank memory macro structures are used. Data are held in a sense amplifier for every memory bank. When access is hit to the held data, data latched by the sense amplifier are output to thereby enhance the speed of first access to the memory macro structures. Namely, each memory bank is made to function as a sense amplifier cache. To enhance the hit ratio of such a sense amplifier cache more greatly, an access controller self-prefetches the next address (an address to which a predetermined offset has been added) after access to a memory macro structure so that data in the self-prefetched address are preread by a sense amplifier in another memory bank.
摘要:
To enhance the speed of first access (read access different in word line from the previous access) to a multi-bank memory, multi-bank memory macro structures are used. Data are held in a sense amplifier for every memory bank. When access is hit to the held data, data latched by the sense amplifier are output to thereby enhance the speed of first access to the memory macro structures. Namely, each memory bank is made to function as a sense amplifier cache. To enhance the hit ratio of such a sense amplifier cache more greatly, an access controller self-prefetches the next address (an address to which a predetermined offset has been added) after access to a memory macro structure so that data in the self-prefetched address are preread by a sense amplifier in another memory bank.
摘要:
To enhance the speed of first access (read access different in word line from the previous access) to a multi-bank memory, multi-bank memory macro structures are used. Data are held in a sense amplifier for every memory bank. When access is hit to the held data, data latched by the sense amplifier are output to thereby enhance the speed of first access to the memory macro structures. Namely, each memory bank is made to function as a sense amplifier cache. To enhance the hit ratio of such a sense amplifier cache more greatly, an access controller self-prefetches the next address (an address to which a predetermined offset has been added) after access to a memory macro structure so that data in the self-prefetched address are preread by a sense amplifier in another memory bank.
摘要:
A dynamic memory requires refreshing to retain data in its memory cells. This may cause access to the dynamic memory for purposes other than refreshing (external access) and access to it for refreshing to compete with each other, resulting in a performance deterioration. According to this invention, a pipelined dynamic memory (PDRAM) is used, and the pipeline frequency (CLK) of the pipelined dynamic memory is made higher than the frequency (CLK1) of external access, and access required for refreshing is made to an unoccupied slot (a timing when any external access request is never issued) in the pipeline of the pipelined dynamic memory. This makes refreshing of the internal dynamic memory an internal operation, which eliminates the need to take refreshing into consideration at the time external access is made, leading to improvement in operating ease and speed.
摘要:
A dynamic memory requires refreshing to retain data in its memory cells. This may cause access to the dynamic memory for purposes other than refreshing (external access) and access to it for refreshing to compete with each other, resulting in a performance deterioration. According to this invention, a pipelined dynamic memory (PDRAM) is used, and the pipeline frequency (CLK) of the pipelined dynamic memory is made higher than the frequency (CLK1) of external access, and access required for refreshing is made to an unoccupied slot (a timing when any external access request is never issued) in the pipeline of the pipelined dynamic memory. This makes refreshing of the internal dynamic memory an internal operation, which eliminates the need to take refreshing into consideration at the time external access is made, leading to improvement in operating ease and speed.