Disque Dur Pour Protools, firewire ou IDE |
|
|
|
|
Replies
(1 - 5)
|
Mon 16 Dec 2002, 11:07
|
Junior Member
Group: Members
Posts: 155
Joined: 25-Aug 02
From: Paris - FR
Member No.: 7,113
|
OFF TOPIC Alors Mr T, ton icone est parfait, ça sent les vacances à plein nez!!!!! Ne change rien c'est top Burns
|
|
|
|
|
Mon 16 Dec 2002, 11:37
|
SuperHero
Group: Members
Posts: 9,465
Joined: 04-Nov 01
From: Paris - FR
Member No.: 2,244
|
Thanks Burns. J'ai effectivement mis le Sombrero pour détendre un peu l'atmosphère... Pour ce qui nous occupe ici, je ne suis personnellement pas un spécialite de l'architecture complexe des disques durs (chacun son métier), mais je crois pouvoir dire que plus la mémoire cache est grande, meilleure ce sera qql soit le type d'utilisation à laquelle tu destine ton HD. Pour les marques: IBM, Seagate, Western Digital, Maxtor...pour moi c'est bonnet blanc, blanc bonnet (d'autres te diront qu'il y a une différence)... Moi j'ai choisi les IBM mais n'en ferais pas plus la publicité qu'aux autres marques. Compare les prix et prends le moins cher! (7200 tr). Voilà. Si tu veux vraiment te pencher sur cette histoire de cache, je t'ais dégoté un "article" trouvé sur le site d'IBM qui traite d'un nouveau disque avec cache "supertopquitue" et qui apporte qqls éléments de réponse sur ce qu'est cette cache (j'espère que tu lis l'anglais) :
"As was previously established, there are existing read cache management designs that involve reading data ahead with the hope that the host system would desire that data. This read-ahead, or pre-fetch, type cache is simplistic and has very little overhead. However, it could be classified as being reactionary. As data is being read from the disks, the pre-fetch cache reacts by reading even more data into cache, anticipating that the next read request would be sequential or near sequential to the previous read. The shortcomings of this type of cache strategy lie in its inability to adapt to a workload, causing, in some cases, decreased performance as it chases and fruitlessly caches data for non-sequential read requests.
On the other hand, there are also cache implementations generally available in the industry that attempt to distinguish data types, classify those data types, and locate each type in a separate region in the cache memory. These complicated caches then attempt to manage each region according to data type. This type of caching strategy can be effective on some workloads, but due to the vast amount of overhead needed to track and classify the data, the performance of the cache itself may suffer."
--------------------
|
|
|
|
|
|
1 User(s) are reading this topic (1 Guests and 0 Anonymous Users)
0 Members:
|
|
|