英语翻译2.3 Last Stream PredictorThe third stage uses a simple last value predictor asshown in Fig.3 to exploit redundancy in the SI tracecomponent caused by repeating sequences of stream indices.A linear predictor table with NP entries is indexe

英语翻译2.3 Last Stream PredictorThe third stage uses a simple last value predictor asshown in Fig.3 to exploit redundancy in the SI tracecomponent caused by repeating sequences of stream indices.A linear predictor table with NP entries is indexe
英语翻译
2.3 Last Stream Predictor
The third stage uses a simple last value predictor as
shown in Fig.3 to exploit redundancy in the SI trace
component caused by repeating sequences of stream indices.
A linear predictor table with NP entries is indexed by
a hash function that is based on the history of previous
stream indices.If the selected predictor entry matches the
incoming stream index,we have an LSP hit.Otherwise,
we have an LSP miss,and the selected predictor entry is
updated with the incoming stream index.The hit/miss
information (one bit,a ‘0’ for a miss and a ‘1’ for a hit)
and,in case of an LSP miss,SI (log2(NSET*NWAYS) bits) are
forwarded to the trace encoder.
The compression ratio achievable by the LSP stage
alone,CR(LSP),can be calculated as shown in (2).It depends
on the stream index size and the LSP hit rate,
hrLSP.The maximum compression ratio that can be
achieved by this stage is log2(NSET*NWAY).The design
space exploration for the last stream predictor includes
determining a good hash function and a number of entries
in the predictor NP.

英语翻译2.3 Last Stream PredictorThe third stage uses a simple last value predictor asshown in Fig.3 to exploit redundancy in the SI tracecomponent caused by repeating sequences of stream indices.A linear predictor table with NP entries is indexe
2.3最后流预测
第三阶段使用一个简单的最后一个值的预测为
如图所示. 3利用冗余跟踪的SI
组件造成的重复序列流指标.
一个线性预测与NP条目索引的表
哈希函数是基于过去的历史基础
流指标.如果选定的预测相匹配的条目
传入的流索引,我们有一个LSP的打击.否则,
我们有一个LSP的小姐,并选定预测条目
更新传入流索引.命中/小姐
资料（1位,一0一错过和'1'的一击）
并错过了一个LSP的情况下,硅（的log2（n设定* NWAYS）位）是
转发到跟踪编码器.
压缩比由LSP阶段实现
仅华润（LSP）的,可在计算如下（2）.这取决于
在流索引的大小和LSP的命中率,
hrLSP.最大压缩比可
到了这个阶段取得的成就的log2（n设定*素内在）.设计
在过去的空间探索包括流预测
确定一个好的散列函数和项目数
在预测镎.