diff --git a/doc/nvme.md b/doc/nvme.md
index 56d8d240e..3565156d7 100644
--- a/doc/nvme.md
+++ b/doc/nvme.md
@@ -58,24 +58,24 @@ demonstrate how to use perf.
 
 Example: Using perf for 4K 100% Random Read workload to a local NVMe SSD for 300 seconds
 ~~~{.sh}
-perf -q 128 -s 4096 -w randread -r 'trtype:PCIe traddr:0000:04:00.0' -t 300
+perf -q 128 -o 4096 -w randread -r 'trtype:PCIe traddr:0000:04:00.0' -t 300
 ~~~
 
 Example: Using perf for 4K 100% Random Read workload to a remote NVMe SSD exported over the network via NVMe-oF
 ~~~{.sh}
-perf -q 128 -s 4096 -w randread -r 'trtype:RDMA adrfam:IPv4 traddr:192.168.100.8 trsvcid:4420' -t 300
+perf -q 128 -o 4096 -w randread -r 'trtype:RDMA adrfam:IPv4 traddr:192.168.100.8 trsvcid:4420' -t 300
 ~~~
 
 Example: Using perf for 4K 70/30 Random Read/Write mix workload to all local NVMe SSDs for 300 seconds
 ~~~{.sh}
-perf -q 128 -s 4096 -w randrw -M 70 -t 300
+perf -q 128 -o 4096 -w randrw -M 70 -t 300
 ~~~
 
 Example: Using perf for extended LBA format CRC guard test to a local NVMe SSD,
 users must write to the SSD before reading the LBA from SSD
 ~~~{.sh}
-perf -q 1 -s 4096 -w write -r 'trtype:PCIe traddr:0000:04:00.0' -t 300 -e 'PRACT=0,PRCKH=GUARD'
-perf -q 1 -s 4096 -w read -r 'trtype:PCIe traddr:0000:04:00.0' -t 200 -e 'PRACT=0,PRCKH=GUARD'
+perf -q 1 -o 4096 -w write -r 'trtype:PCIe traddr:0000:04:00.0' -t 300 -e 'PRACT=0,PRCKH=GUARD'
+perf -q 1 -o 4096 -w read -r 'trtype:PCIe traddr:0000:04:00.0' -t 200 -e 'PRACT=0,PRCKH=GUARD'
 ~~~
 
 # Public Interface {#nvme_interface}
@@ -227,8 +227,8 @@ Example: identical shm_id and non-overlapping core masks
 	[-c core mask for I/O submission/completion]
 	[-i shared memory group ID]
 
-./perf -q 1 -s 4096 -w randread -c 0x1 -t 60 -i 1
-./perf -q 8 -s 131072 -w write -c 0x10 -t 60 -i 1
+./perf -q 1 -o 4096 -w randread -c 0x1 -t 60 -i 1
+./perf -q 8 -o 131072 -w write -c 0x10 -t 60 -i 1
 ~~~
 
 ## Limitations {#nvme_multi_process_limitations}
diff --git a/examples/ioat/perf/perf.c b/examples/ioat/perf/perf.c
index d98b7a079..fdc222096 100644
--- a/examples/ioat/perf/perf.c
+++ b/examples/ioat/perf/perf.c
@@ -250,7 +250,7 @@ usage(char *program_name)
 	printf("\t[-c core mask for distributing I/O submission/completion work]\n");
 	printf("\t[-q queue depth]\n");
 	printf("\t[-n number of channels]\n");
-	printf("\t[-s transfer size in bytes]\n");
+	printf("\t[-o transfer size in bytes]\n");
 	printf("\t[-t time in seconds]\n");
 	printf("\t[-v verify copy result if this switch is on]\n");
 }
@@ -261,9 +261,9 @@ parse_args(int argc, char **argv)
 	int op;
 
 	construct_user_config(&g_user_config);
-	while ((op = getopt(argc, argv, "c:hn:q:s:t:v")) != -1) {
+	while ((op = getopt(argc, argv, "c:hn:o:q:t:v")) != -1) {
 		switch (op) {
-		case 's':
+		case 'o':
 			g_user_config.xfer_size_bytes = atoi(optarg);
 			break;
 		case 'n':
diff --git a/examples/nvme/perf/perf.c b/examples/nvme/perf/perf.c
index e7a8ec4ad..b5418b473 100644
--- a/examples/nvme/perf/perf.c
+++ b/examples/nvme/perf/perf.c
@@ -911,7 +911,7 @@ static void usage(char *program_name)
 #endif
 	printf("\n");
 	printf("\t[-q io depth]\n");
-	printf("\t[-s io size in bytes]\n");
+	printf("\t[-o io size in bytes]\n");
 	printf("\t[-w io pattern type, must be one of\n");
 	printf("\t\t(read, write, randread, randwrite, rw, randrw)]\n");
 	printf("\t[-M rwmixread (100 for reads, 0 for writes)]\n");
@@ -938,7 +938,7 @@ static void usage(char *program_name)
 	printf("\t  PRCHK      Control of Protection Information Checking (PRCHK=GUARD|REFTAG|APPTAG)\n");
 	printf("\t Example: -e 'PRACT=0,PRCHK=GUARD|REFTAG|APPTAG'\n");
 	printf("\t          -e 'PRACT=1,PRCHK=GUARD'\n");
-	printf("\t[-d DPDK huge memory size in MB.]\n");
+	printf("\t[-s DPDK huge memory size in MB.]\n");
 	printf("\t[-m max completions per poll]\n");
 	printf("\t\t(default: 0 - unlimited)\n");
 	printf("\t[-i shared memory group ID]\n");
@@ -1280,14 +1280,11 @@ parse_args(int argc, char **argv)
 	g_core_mask = NULL;
 	g_max_completions = 0;
 
-	while ((op = getopt(argc, argv, "c:d:e:i:lm:q:r:s:t:w:DLM:")) != -1) {
+	while ((op = getopt(argc, argv, "c:e:i:lm:o:q:r:s:t:w:DLM:")) != -1) {
 		switch (op) {
 		case 'c':
 			g_core_mask = optarg;
 			break;
-		case 'd':
-			g_dpdk_mem = atoi(optarg);
-			break;
 		case 'e':
 			if (parse_metadata(optarg)) {
 				usage(argv[0]);
@@ -1303,6 +1300,9 @@ parse_args(int argc, char **argv)
 		case 'm':
 			g_max_completions = atoi(optarg);
 			break;
+		case 'o':
+			g_io_size_bytes = atoi(optarg);
+			break;
 		case 'q':
 			g_queue_depth = atoi(optarg);
 			break;
@@ -1313,7 +1313,7 @@ parse_args(int argc, char **argv)
 			}
 			break;
 		case 's':
-			g_io_size_bytes = atoi(optarg);
+			g_dpdk_mem = atoi(optarg);
 			break;
 		case 't':
 			g_time_in_sec = atoi(optarg);
diff --git a/test/nvme/nvme.sh b/test/nvme/nvme.sh
index 18b8800f5..9cdcb21e2 100755
--- a/test/nvme/nvme.sh
+++ b/test/nvme/nvme.sh
@@ -111,10 +111,10 @@ done
 timing_exit identify
 
 timing_enter perf
-$rootdir/examples/nvme/perf/perf -q 128 -w read -s 12288 -t 1 -LL -i 0
+$rootdir/examples/nvme/perf/perf -q 128 -w read -o 12288 -t 1 -LL -i 0
 if [ -b /dev/ram0 ]; then
 	# Test perf with AIO device
-	$rootdir/examples/nvme/perf/perf /dev/ram0 -q 128 -w read -s 12288 -t 1 -LL -i 0
+	$rootdir/examples/nvme/perf/perf /dev/ram0 -q 128 -w read -o 12288 -t 1 -LL -i 0
 	report_test_completion "nvme_perf"
 fi
 timing_exit perf
@@ -153,11 +153,11 @@ timing_exit arbitration
 
 if [ `uname` = Linux ]; then
 	timing_enter multi_secondary
-	$rootdir/examples/nvme/perf/perf -i 0 -q 16 -w read -s 4096 -t 3 -c 0x1 &
+	$rootdir/examples/nvme/perf/perf -i 0 -q 16 -w read -o 4096 -t 3 -c 0x1 &
 	pid0=$!
-	$rootdir/examples/nvme/perf/perf -i 0 -q 16 -w read -s 4096 -t 3 -c 0x2 &
+	$rootdir/examples/nvme/perf/perf -i 0 -q 16 -w read -o 4096 -t 3 -c 0x2 &
 	pid1=$!
-	$rootdir/examples/nvme/perf/perf -i 0 -q 16 -w read -s 4096 -t 3 -c 0x4
+	$rootdir/examples/nvme/perf/perf -i 0 -q 16 -w read -o 4096 -t 3 -c 0x4
 	wait $pid0
 	wait $pid1
 	report_test_completion "nvme_multi_secondary"
diff --git a/test/nvmf/host/perf.sh b/test/nvmf/host/perf.sh
index e39d6e2f3..ef5214c8e 100755
--- a/test/nvmf/host/perf.sh
+++ b/test/nvmf/host/perf.sh
@@ -44,10 +44,10 @@ $rpc_py construct_nvmf_subsystem nqn.2016-06.io.spdk:cnode1 "trtype:RDMA traddr:
 
 # Test multi-process access to local NVMe device
 if [ -n "$local_nvme_trid" ]; then
-	$rootdir/examples/nvme/perf/perf -i 0 -q 32 -s 4096 -w randrw -M 50 -t 1 -r "$local_nvme_trid"
+	$rootdir/examples/nvme/perf/perf -i 0 -q 32 -o 4096 -w randrw -M 50 -t 1 -r "$local_nvme_trid"
 fi
 
-$rootdir/examples/nvme/perf/perf -q 32 -s 4096 -w randrw -M 50 -t 1 -r "trtype:RDMA adrfam:IPv4 traddr:$NVMF_FIRST_TARGET_IP trsvcid:4420"
+$rootdir/examples/nvme/perf/perf -q 32 -o 4096 -w randrw -M 50 -t 1 -r "trtype:RDMA adrfam:IPv4 traddr:$NVMF_FIRST_TARGET_IP trsvcid:4420"
 sync
 $rpc_py delete_nvmf_subsystem nqn.2016-06.io.spdk:cnode1
 
@@ -63,8 +63,8 @@ if [ $RUN_NIGHTLY -eq 1 ]; then
 		qd_depth=("1" "128")
 		io_size=("512" "131072")
 		for qd in ${qd_depth[@]}; do
-			for s in ${io_size[@]}; do
-				$rootdir/examples/nvme/perf/perf -q $qd -s $s -w randrw -M 50 -t 10 -r "trtype:RDMA adrfam:IPv4 traddr:$NVMF_FIRST_TARGET_IP trsvcid:4420"
+			for o in ${io_size[@]}; do
+				$rootdir/examples/nvme/perf/perf -q $qd -o $o -w randrw -M 50 -t 10 -r "trtype:RDMA adrfam:IPv4 traddr:$NVMF_FIRST_TARGET_IP trsvcid:4420"
 			done
 		done