Breaking News
News & Analysis

Atom gets first ride in HP's CPU-agnostic server

6/19/2012 05:57 PM EDT
4 comments
NO RATINGS
More Related Links
View Comments: Oldest First | Newest First | Threaded View
chanj0
User Rank
Author
re: Atom gets first ride in HP's CPU-agnostic server
chanj0   6/19/2012 7:59:38 PM
NO RATINGS
300 concurrent web sessions seem a bit low to me. How much RAM does the server equip? When the comparison is done, is the 150W Xeon powered server run only 300 concurrent web sessions?

rick merritt
User Rank
Author
re: Atom gets first ride in HP's CPU-agnostic server
rick merritt   6/19/2012 8:15:38 PM
NO RATINGS
Yes the comparison is based on similar workloads on the two processors, but HP is not commenting on how much memory it uses or any other details of the Gemini chassis or CPU cartridges.

green_is_now
User Rank
Author
re: Atom gets first ride in HP's CPU-agnostic server
green_is_now   6/24/2012 3:26:20 PM
NO RATINGS
Looked at the Calxeda quad, quad core server card or energy card, 16 A9's, 24Watts. I would like to see someone with insight talk to what runs well on ARM SISP and what runs well on CISP. What is a software bottleneck and what is a hardware bottleneck. Both single user and server farm levels. Can this agnostic approuch from HP support both in the same server cluster parsing out CISP optimized workloads and SISP optimized workloads. How does cload computing change things when the application software is at the server and the results are piped to the user. Does this change the % need of the two. Obvuiosly the lowestcost operation will prevail for most all tasks that can be done with Sisp. Would not more cloud activity tilt the field toward SISP or would it require more Cisp machines? Does virtualisation and othe things untether apps form cisp?

selinz
User Rank
Author
re: Atom gets first ride in HP's CPU-agnostic server
selinz   12/23/2013 2:22:35 PM
NO RATINGS
Intel has the critical mass to force down their margins to whatever it takes to win market share. In this economy, it's tough to be a little guy.

Most Recent Comments
michigan0
 
SteveHarris0
 
realjjj
 
SteveHarris0
 
SteveHarris0
 
VicVat
 
Les_Slater
 
SSDWEM
 
witeken
Most Recent Messages
9/25/2016
4:48:30 PM
michigan0 Sang Kim First, 28nm bulk is in volume manufacturing for several years by the major semiconductor companies but not 28nm FDSOI today yet. Why not? Simply because unlike 28nm bulk the LDD(Lightly Doped Drain) to minimize hot carrier generation can't be implemented in 28nm FDSOI. Furthermore, hot carrier reliability becomes worse with scaling, That is the major reason why 28nm FDSOI is not manufacturable today and will not be. Second, how can you suppress the leakage currents from such ultra short 7nm due to the short channel effects? How thin SOI thickness is required to prevent punch-through of un-dopped 7nm FDSOI? Possibly less than 4nm. Depositing such an ultra thin film less then 4nm filum uniformly and reliably over 12" wafers at the manufacturing line is extremely difficult or not even manufacturable. If not manufacturable, the 7nm FDSOI debate is over!Third, what happens when hot carriers are generated near the drain at normal operation of 7nm FDSOI? Electrons go to the positively biased drain with no harm but where the holes to go? The holes can't go to the substrate because of the thin BOX layer. Some holes may become trapped at the BOX layer causing Vt shift. However, the vast majority of holes drift through the the un-dopped SOI channel toward the N+Source,...

Datasheets.com Parts Search

185 million searchable parts
(please enter a part number or hit search to begin)
Like Us on Facebook
EE Times on Twitter
EE Times Twitter Feed