Viewing Analysis Results

Prerequisites

An overall analysis task has been created and the analysis is complete.

Procedure

In the System Profiler area on the left, click the name of the target analysis task.
The node list is displayed.

Click the name of the target node to view the analysis results. You can click

in the upper right corner to download the overall analysis data.

Click the node name. The Summary tab page is displayed by default, as shown in Figure 1. For details about the parameters, see the following tables.

The Summary tab page displays the information about the CPU, memory subsystem, storage subsystem, network subsystem, and other configuration items. If the physical machine environment is used, the system component diagram is displayed.
The Tuning Suggestions area is displayed only when the system detects metrics to be tuned. You can expand the area to view detailed suggestions and solutions. Tuning suggestions are provided for certain configuration items and marked with . Certain parameters are marked with . You can move the mouse pointer onto the icon to view detailed parameter description and tuning suggestions.

Figure 1 System Settings

**Table 1** Parameters in the CPU Package area
Parameter	Description
CPU Model	Model of the CPU.
Cores	Number of CPU cores.
Max. Frequency	Maximum frequency of the CPU.
Current Frequency	Current frequency of the CPU.
CPU Core
ID	CPU manufacturer ID. For example, 0x48 indicates HiSilicon.
Max. Frequency	Maximum frequency of the CPU.
Current Frequency	Current frequency of the CPU.
L1I	Size of the L1 high-speed instruction cache of the CPU.
L2	Size of the L2 cache of the CPU.
L1D	Size of the L1 high-speed data cache of the CPU.
L3 cache
Cache Size	Size of the L3 cache of the CPU.
Nodes Sharing L3 Cache	NUMA nodes that share the L3 cache.
NUMA Nodes
Node	NUMA node name.
CPU Core ID	CPU core ID of the NUMA node.
Total Memory	Total memory size of the NUMA node.
Idle Memory	Idle memory size of the NUMA node.
NUMA Node Distance
Node	NUMA node name. NOTE: The distance between NUMA nodes represents the costs of a NUMA node to access the memory of another node. A shorter distance means faster memory access from the CPU.
NUMA Balancing
NUMA Balancing	Setting of the NUMA balancing function.

**Table 2** Parameters in the GPU Package area
Parameter	Description
GPU ID	GPU ID.
Driver Version	GPU driver version.
CUDA Version	CUDA version.
Product Name	Product name.
Product Brand	Product brand.
Display Mode	Display mode flag.
Display Active	Display active flag.
Persistence Mode	Persistence mode flag.
Bus ID	PCIe bus ID.
Max PCIe Generation	Maximum PCIe generation.
Max PCIe Link Width	Maximum PCIe bandwidth.
VBIOS Version	Onboard BIOS version.
FB Memory Size (MB)	Memory capacity of the device.
BAR1 Memory Size (MB)	Size of the mapped memory.
Graphics Clock (MHz)	Graphics clock frequency.
Memory Clock (MHz)	Memory clock frequency.
Compute Mode	Compute mode.
Ecc Mode	Error check mode.

**Table 3** Parameters in the Memory Subsystem area
Parameter	Description
Total Memory Capacity	Total memory capacity of the system.
DIMMs	Number of DIMMs.
Empty Slots	Number of empty memory slots.
DIMM List
Slot	Position of the DIMM slot.
Capacity	Capacity of the DIMM installed in the current slot.
Max. Rate	Maximum memory speed.
Configured Rate	Memory speed configured.
Type	Type of the DIMM.

**Table 4** Parameters in the Storage Subsystem area
Parameter	Description
Total Capacity	Total storage capacity.
Total Drives	Total number of drives.
Drives
Device Name	Device name.
Drive Model	Model of the drive.
Drive Capacity	Capacity of the drive.
Drive Type	Type of the drive.
RAID Controller Cards
Controller Model	Model of the RAID controller card.
Card ID	ID of the RAID controller card.
Cache Size	Size of the cache.

**Table 5** Parameters in the Network Subsystem area
Parameter	Description
NICs	Number of NICs.
Network Ports	Number of network ports.
NICs
Latency	Latency of the NIC.
NUMA Node	NUMA node name.
Kernel Driver	Driver of the kernel.
Kernel Module	Kernel module.
Subsystem	Subsystem.
Network Ports
Port	Name of the network port.
Device (B/D/F)	Bus/Device/Function (B/F/D) of the device.
Hardware ID	Hardware ID.
txkB/s	Total data transmitted per second, in KB.
rxkB/s	Total data received per second, in KB.
txpck/s	Total number of packets transmitted per second.
rxpck/s	Total number of packets received per second.
Max. Transmission Rate	Maximum transmission speed.
Current Configured Rate	Current transmission speed.
Max. Data Load (byte)	Maximum data load.
Details	Details about the device.

**Table 6** Parameters in the Runtime Environment Information area
Parameter	Description
Basic System Information
BIOS	BIOS version information.
OS Version	OS version information.
Kernel Version	System kernel version information.
JDK Version	Java Development Kit (JDK) version information.
glibc Version	glibc version information.
system_dmesg	System dmesg information. Click View Details for details.
docker info	Docker container information. Click View Details for details. NOTE: This parameter is displayed only when a Docker container is installed on the host machine.
sysctl	Information about all system control configuration items. Click View Details for details.
kernelConfig	Kernel configuration information. Click View Details for details.
docker images	Docker container image information. Click View Details for details. NOTE: This parameter is displayed only when a Docker container is installed on the host machine.
BMC Firmware Version	Firmware version of the BMC.
SSD driver
Driver Version	Driver version.
SSD Firmware Version	Firmware version of the SSD.
Memory Management System
SMMU	Status of the SMMU function.
PT Size (byte)	Size of the page table.
THP	Status of the transparent huge page (THP) function.
Standard Hugepage (KB)	Size of the standard huge page.
Number of Huge Pages	Number of the standard huge pages. The value 0 indicates that this parameter is not set.
Swap Partition	Size of the current swap partition.
Remaining Cache Time of Dirty Data (centisecond)	Remaining time of the dirty data in the cache.
Max. Ratio of Dirty Pages to Total Memory	Maximum ratio of the dirty page size to the total memory size.
Max. Ratio of Dirty Page Cache to Total Memory	Maximum ratio of the dirty page cache size to the total memory size.
Interval for Waking Up the pdflush Process to Refresh Dirty Data	Interval for waking up the pdflush process to refresh dirty data, in centiseconds.
Minimum Free Memory Reserved (KB)	Minimum size of the idle memory reserved, in KB.
NIC Firmware Version	NIC port and NIC firmware version.
VMs/Dockers
VM Libvirt Version	VM libvirt version.
KVM VM Configuration	Kernel-based Virtual Machine (KVM) configuration parameters.
Docker Version	Docker version.
Kernel Parameters
HZ Value	HZ value.
NOHZ Timer	Status of the NOHZ timer.
cmdline	Startup script of the kernel.

**Table 7** Parameters in the Storage Resource Configuration area
Parameter	Description
Volumes	Number of volumes.
File Systems	Number of file system partitions.
RAID Arrays	Number of RAID arrays.
RAID level
Volume Name	Logical drive name.
Volume ID	Logical drive ID.
RAID Controller Card ID	ID of the RAID controller card.
RAID Level	RAID level.
Logical Drive Strip Size	Size of the logical drive strip.
Current Logical Drive Read Policy	Current read policy of the logical drive.
Current Logical Drive Write Policy	Current write policy of the logical drive.
Current Logical Drive Cache Policy	Current cache policy of the local drive.
CacheCade Identifier	CacheCade identifier.
RAID Configuration	Current RAID configuration.
Storage Info
Device Name	Device name.
Prefetch Size of Drive Files (KB)	Prefetch size of the files on drives.
Storage I/O Scheduling Mechanism	I/O scheduling mechanism of the storage.
Drive Request Affinity Setting	Setting of the drive request affinity. The value 1 indicates that the I/O request is completed by the CPU that initiates the I/O request.
Drive Request Queue Length	Average number of read and write requests for a specified drive in a queue within a sampling interval.
Drive Queue Depth	Number of I/O operations that can be concurrently processed by the device when the host initiates an I/O request.
I/O Merging	I/O merging setting. 0: enables all types of merge attempts. 1: disables complex merge checks, but leaves the simple merge with the previous I/O request valid. 2: disables all types of merge attempts.
File System Info
Partition	Partition name.
File System Type	File system type of the partition.
Mount Point	Mount point of the current partition.
Mount Information	Mount information of the current partition.

**Table 8** Parameters in the Network Port Configuration area
Parameter	Description
Ports	Number of network ports.
Interrupt Aggregation
Network Port Name	Name of the network port.
adaptive-rx	Setting of the dynamic aggregation for a receiving queue.
adaptive-tx	Setting of the dynamic aggregation for a transmit queue.
rx-usecs	Number of microseconds after at least one packet is received before an interrupt is generated.
tx-usecs	Number of microseconds after at least one packet is transmitted before an interrupt is generated.
rx-frames	Number of packets transmitted before an interrupt is generated.
tx-frames	Number of packets received before an interrupt is generated.
Offload
Network Port Name	Name of the network port.
rx-checksumming	Setting of the checksum function for received packets.
tx-checksumming	Setting of the checksum function for transmitted packets.
scatter-gather	Setting of the scatter/gather function.
TSO	Setting of the TCP segmentation offload (TSO) function.
UFO	Setting of the UDP fragmentation offload (UFO) function.
LRO	Setting of the large receive offload (LRO) function.
GSO	Setting of the generic segmentation offload (GSO) function.
GRO	Setting of the generic receive offload (GRO) function.
Queue
Network Port Name	Name of the network port.
Queues	Number of queues.
Interrupt NUMA Core Pinning
Network Port Name	Name of the network port.
Interrupt ID	Interrupt ID.
Interrupt NUMA Core Pinning Info	Information about the pinning of interrupts and NUMA cores.
xps/rps	Core pinning information of the transmit or receive queue.
Ring Buffer
NIC Name	NIC name.
TX (Byte)	Size of the sent cache ring buffer, in bytes.
RX (Byte)	Size of the received cache ring buffer, in bytes.

Click PCIe Topology to view the PCIe topology analysis result, as shown in Figure 2. Table 9 describes the parameters.

You can move the mouse pointer to a device icon in the topology view to view the device details and suggestions. indicates a NUMA node, indicates the interface between the CPU and PCIe topology, indicates the PCI bridge, and indicates the PCIe device.
Click + or - to zoom in or zoom out the topology view.
You can move the mouse pointer to the icon to view the parameter details.
In the interrupt distribution statistics area, you can click to filter the hardware interrupts to be viewed. You can select interrupts from the drop-down list.
The icon next to the displayed version information indicates that a new driver version of the NIC is available for upgrade. You can click the icon to perform upgrade operations.

Figure 2 PCIe topology information

The icon next to the displayed version information indicates that a new driver version of the NIC is available for upgrade. You can click the icon to perform upgrade operations.

**Table 9** Parameters on the PCIe Topology tab page (NIC device information)
Parameter	Description
PCIe device information
device_full_name	Device name.
DevCap	Capability of the device.
DevCtl	Device scheduling information.
LnkCap	Link capability.
LnkCtl	Link scheduling information.
LnkSta	Link status.
Device Firmware and Driver Information
driver	Driver name.
version	Driver version.
firmware-version	Driver firmware information.
driver mount params	Driver parameter information.
supported-ports	Information about supported ports.
Network Port Status Info
Network Name	Name of a network device.
Link detected	Link status of the NIC.
Speed	NIC speed.
Duplex	Work mode of the NIC.
Auto-negotiation	Indicates whether auto-negotiation is enabled for the NIC.
Port Statistics
rxbyte/s	Number of bytes received per second.
txbyte/s	Number of bytes transmitted per second.
rxdrop/s	Number of received data packets that are discarded per second.
txdrop/s	Number of transmitted data packets that are discarded per second.
Receive bytes	Total number of bytes received by the interface.
Transmit bytes	Total number of bytes transmitted by the interface.
Receive drop	Total number of discarded data packets received by the driver.
Transmit drop	Total number of discarded data packets transmitted by the driver.
net_core_statistic	Net core statistics.
driver_statistic	Driver layer statistics.
driver_queue_statistic	Statistics about queues at the driver layer.
firmware_statistic	Microcode-layer statistics.
MAC_statistic	Physical layer statistics.
Interrupt Request Configuration
irqbalance	Balance status of an interrupt request.
RPS	Indicates whether Receive Packet Steering (RPS) is supported.
XPS	Indicates whether Transmit Packet Steering (XPS) is supported.
RFS	Indicates whether Receive Flow Steering (RFS) is supported.
aRFS	Indicates whether accelerated RFS (aRFS) is supported.
Interrupt binding core settings	Number of cores bound to the interrupt.
Configuration information about the queue size and ring buffer size
RSS	Indicates whether Receive Side Scaling (RSS) is supported.
Queue Configuration	Queue configuration.
ring buffer Configuration	Ring buffer configuration.
Configuration Information of the NIC Offload Function
rx-checksumming	Indicates whether the hardware checksum calculation on the receive end (RX) is supported.
tx-checksumming	Indicates whether the hardware checksum calculation on the transmit end (TX) is supported.
tx-checksum-ipv4	Indicates whether the IPv4-based TCP/UDP hardware checksum calculation on the TX is supported.
tx-checksum-ip-generic	Indicates whether the IP checksum calculation of any protocol or protocol layer combination on the TX is supported.
tx-checksum-ipv6	Indicates whether the IPv6-based TCP/UDP hardware checksum calculation on the TX is supported.
tx-checksum-fcoe-crc	Indicates whether the hardware checksum calculation of the FCoE CRC32 is supported.
tx-checksum-sctp	Indicates whether the SCTP hardware checksum calculation on the TX is supported.
scatter-gather	Indicates whether scatter-gather is enabled.
tx-scatter-gather	Indicates whether scatter-gather is enabled. ndo_start_xmit can process page-based fragmented skbs(skb_shinfo()->frags).
tx-scatter-gather-fraglist	Indicates whether scatter-gather is enabled. ndo_start_xmit can process chained fragmented skbs(skb->next/prev list).
tcp-segmentation-offload	Indicates whether TSO is enabled.
tx-tcp-segmentation	Indicates whether TCPv4 segmentation is enabled.
tx-tcp-ecn-segmentation	Indicates whether hardware is enabled to correctly split data packets with the CWR bit for TCPv4 (tx-tcp-segmentation must be enabled) or TCPv6 (tx-tcp6-segmentation must be enabled).
tx-tcp-mangleid-segmentation	Indicates whether the IP ID can be ignored. (The IP ID can increase with the segmentation or remain unchanged, based on the preferred settings of the drive.)
tx-tcp6-segmentation	Indicates whether TCPv6 segmentation is enabled.
generic-segmentation-offload	Indicates whether segmentation offload is enabled.
generic-receive-offload	Indicates whether GRO is enabled.
large-receive-offload	Indicates whether LRO is enabled.
rx-vlan-offload	Indicates whether VLAN CTAG hardware acceleration on the RX is supported.
tx-vlan-offload	Indicates whether VLAN CTAG hardware acceleration on the TX is supported.
receive-hashing	Indicates whether hashing offload on the RX is enabled.
rx-vlan-filter	Indicates whether VLAN CTAGs filtering on the RX is enabled.
tx-fcoe-segmentation	Indicates whether FCoE segmentation is enabled.
tx-gre-segmentation	Indicates whether the TX segmentation offload of the GRE packet headers is enabled.
tx-gre-csum-segmentation	Indicates whether the TX segmentation offload with checksum in the GRE packet headers is supported.
tx-ipxip4-segmentation	Indicates whether TX segmentation offload of the IP4 or IP6 over IP4 packet header is supported.
tx-ipxip6-segmentation	Indicates whether TX segmentation offload of the IP4 or IP6 over IP6 packet header is supported.
tx-udp_tnl-segmentation	Indicates whether the TX segmentation offload of the UDP tunnel packet headers is supported.
tx-udp_tnl-csum-segmentation	Indicates whether the TX segmentation offload with checksum in the UDP tunnel packet headers is supported.
tx-gso-partial	Indicates whether partial generic segmentation offload is supported.
tx-sctp-segmentation	Indicates whether the function of transmitting segments using SCTP packets is enabled.
tx-esp-segmentation	Indicates whether the function of transmitting segments using ESP packets is enabled.
tx-udp-segmentation	Indicates whether the function of transmitting segments using UDP packets is enabled.
tx-vlan-stag-hw-insert	Indicates whether hardware acceleration for transmitting VLAN S-Tag packets is enabled.
rx-vlan-stag-hw-parse	Indicates whether hardware acceleration for receiving VLAN S-Tag packets is enabled.
rx-vlan-stag-filter	Indicates whether filtering for receiving VLAN S-Tag packets is enabled.
l2-fwd-offload	Indicates whether Layer-2 data packets can be forwarded in hardware.
hw-tc-offload	Indicates whether hardware offload for traffic classifier (TC) is enabled.
esp-hw-offload	Indicates whether hardware offload for ESP packets on the TX is enabled.
esp-tx-csum-hw-offload	Indicates whether ESP packet transmission checksum offload is enabled.
rx-udp_tunnel-port-offload	Indicates whether UDP tunnel offload is enabled on the RX.
tls-hw-tx-offload	Indicates whether per-packet encryption is enabled for the NIC on the RX.
tls-hw-rx-offload	Indicates whether per-packet encryption is enabled for the NIC on the TX.
rx-gro-hw	Indicates whether hardware GRO is enabled.
tls-hw-record	Indicates whether the NIC driver and firmware TCP processing can be used to replace the kernel network stack processing. This function is unavailable when the Linux network stack is used, such as the firewall function, QoS, and data packet scheduling.
Other network configuration information
mtu	Maximum size of data packets that can pass through the NIC.
Pause parameters	Pause frame configuration parameters.
Sysctl variables	Settings of sysctl variables.
LRO coalesce time	Maximum time used for aggregating received packets.
Network Port
hca_id	Network adapter ID.
transport	Protocol information.
hw_ver	Hardware version.
manufacture	Network adapter manufacturer.
node_guid	Unique identifier of a network adapter.
phys_prot_cnt	Number of physical network ports.
port	Number of ports.
state	NIC status.
active_mtu	Current maximum transmission unit (MTU).
sm_lid	LID of the sub-network manager.
port_lid	LID of the port.
port_lmc	Local multicast address of the port.
link_layer	Current working mode.
roce_enable	Indicates whether RoCE is enabled.
active_speed	Current speed.
Ethernet	Ethernet adapter information.
bond	Bond mode information.

**Table 10** Parameters on the PCIe Topology tab page (NVMe device information)
Parameter	Description
PCIe device information
device_full_name	Device name.
DevCap	Capability of the device.
DevCtl	Device scheduling information.
LnkCap	Link capability.
LnkCtl	Link scheduling information.
LnkSta	Link status.
Smart configuration
critical warning	Critical warning information.
available spare	Redundant space proportion.
available spare threshold	Redundant space alarm proportion.
percentage used	Used service life.
data units read	Volume of read data.
data units written	Volume of written data.
host read commands	Total number of host read times.
host write commands	Total number of host write times.
controller busy time	Busy time statistics of the controller.
power cycles	Number of power-on times.
power on hours	Power-on duration.
unsafe shutdowns	Number of unsafe shutdown times.
media and data integrity errors	Number of data errors.
number of error information log entries	Number of error logs.
warning composite temperature time	Time of minor warning about composite temperature.
critical composite temperature time	Time of major warning about composite temperature.
data status	Data health status.
Basic configuration information
Namespace	Namespace.
maximum capacity	Original maximum capacity.
current capacity	Current maximum capacity.
volatile write cache	Indicates whether to support the write cache.
model number	Device model.
firmware version	Firmware version.
NVMe version	NVMe version.
device status	Device health status.
Advanced settings
controller id	Controller ID.
PCI slot	PCI slot number.
highest temperature	Highest historical temperature.
capacitance status	Supercapacitor health status.
atomic write	Atomic write status.
bad block rate	Bad block rate.
total blocks count	Total number of blocks.
bad blocks count	Total number of bad blocks.
factory bad blocks count	Number of bad blocks before delivery.
flash program failed count	Number of flash programming failures.
flash erase failed count	Number of flash erase failures.
maximum data transfer size	Maximum size of data that can be transmitted.
System configuration information
options nvme use_threaded_interrupts	Option for NVMe to use thread interrupts.
dirty_ratio	Percentage of the dirty data that triggers synchronization to the available memory.
dirty_background_ratio	Percentage of the dirty data amount that triggers flushing to the available memory.
dirty_writeback_centisecs	Cycle of pdflush dirty data writing.
dirty_expire_centisecs	Duration when dirty data resides in the memory.
nvme1	Device details.
Device temperature information
composite temperature	Temperature information.
exceed temperature threshold count	Number of times that the temperature exceeds the threshold.
exceed temperature threshold time	Time when the temperature exceeds the threshold.
composite highest temperature	Historical highest temperature of the entire drive.
composite over temperature threshold count	Overtemperature count of the entire drive.
composite over temperature threshold time	Overtemperature time of the entire drive.
composite current temperature	Current temperature of the entire drive.
composite over temperature threshold temperature	High temperature threshold of the entire drive.
composite under temperature threshold temperature	Low temperature threshold of the entire drive.
controller highest temperature	Historical highest temperature of the controller.
controller over temperature threshold count	Number of controller overtemperature events.
controller over temperature threshold time	Controller overtemperature duration.
controller1 current temperature	Current temperature of the controller.
controller1 over temperature threshold temperature	High temperature threshold of the controller.
controller1 under temperature threshold temperature	Low temperature threshold of the controller.
Other configuration information
power states	Power consumption level of the device.
Write mode	Value of the write mode.
self-test	Diagnosis test result.
disk writes per day	Average amount of data written to the drive per day.
Device I/O statistics
Range (µs)	I/O latency range.
Read I/O Frequency (times/s)	Read I/O frequency.
Write I/O Frequency (times/s)	Write I/O frequency.

Click the System Performance tab page. The performance metrics analysis result is displayed, as shown in Figure 3. For details about the parameters, see the following tables.
- The sequence data of each performance metric in the entire collection process is displayed. You can drag the time bar on the chart to view the time sequence data in a specified period.
- The trends of CPU usage, CPU load, CPU level 1 top-down events, memory usage statistics, storage I/O, network I/O, and energy consumption in the collection period are displayed. You can select System, Numa, or Core from the Type drop-down list to switch the CPU usage dimension.
- You can click next to a system parameter to expand or collapse the corresponding display bar, drag in the display bar to zoom in or zoom out the line chart display range of the corresponding parameter, and click to restore the corresponding line chart to the initial display status.
- In the CPU level 1 top-down data display area, you can click to filter data.
Figure 3 System Performance tab page

Click the Task Scheduling tab to view the system task scheduling information.

The sequence data of system task scheduling information in the entire collection process is displayed. You can drag the time bar on the chart to view the time sequence data in a specified period.
You can click next to a system parameter to expand or collapse the corresponding display bar, drag in the display bar to zoom in or zoom out the line chart display range of the corresponding parameter, and click to restore all line charts to the initial display status.
For top 10 task statistics and perf statistics, you can click on the right of the table header to filter data.
For system task scheduling statistics, you can select Systems, Numa, or Core from the Type drop-down list for data display. For top 10 task statistics, you can select User-mode process or Kernel-mode process from the drop-down list for data display.

Figure 4 Task scheduling

**Table 11** Perf stat Statistics parameters
Parameter	Description
PMU Statistics	Name of the PMU event. context-switches: number of context switches, indicating the number of times that the CPU switches from one process to another. Frequent switches may deteriorate the performance. cpu-migrations: number of CPU migrations, that is, the number of times that a process is scheduled to different CPU cores. This may cause cache failures. page-faults: number of page faults, including minor and major page faults, which deteriorate memory access performance. branch-misses: number of branch prediction failures, which affects the pipeline efficiency. A high failure rate may cause performance deterioration. IPC: number of instructions completed in each clock cycle, which reflects the CPU pipeline efficiency. A high IPC value (for example, > 1.0) indicates high instruction-level parallelism. A low IPC value may be caused by cache invalidation, dependency stagnation, or resource contention.
Average	Average number of each PMU event in the sampling period.

Click the Task Information tab to view the detailed configuration and sampling information about the task on the current node.

If the task fails to be executed, the failure cause is displayed on the Task Information tab page.

If part of data fails to be collected but the overall task execution is not affected, you can view the exception message in Exception Information.

Collection End Cause displays the reason why the data collection of the current task ends, for example, "Task collection times up" or "File size reaches the collection limit."

Parent topic: Overall Analysis